Add Shardok-side connection status indicator

The strategic-map status widget already reports Your Turn / Waiting for Players / etc. via ServerGameStatus. This adds the equivalent for tactical battles so the player knows whether they're up, the AI is thinking, or another human is about to act. ShardokServerStatus is a separate proto from ServerGameStatus so the strategic and tactical layers don't have to evolve in lockstep. The server populates it per-battle in GameController by checking shardokPlayerAvailableCommands across factions, classified into AI vs human via aiClientFactionIds. Client side: ShardokGameModel implements IShardokGameStateProvider and caches the latest status. EagleGameController binds the persistent ConnectionStatusUI to the Shardok model on entering battle; ShardokGameController clears it on exit. Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
2026-07-29 02:48:40 +00:00 · 2026-04-29 21:46:52 -07:00
1836 changed files with 313552 additions and 428829 deletions
@@ -1,4 +1,4 @@
-# bazel-1.0.0.bazelrc
+bazel-1.0.0.bazelrc

 # for now: filter out annoying TASTY warnings
 common --ui_event_filters=-INFO
@@ -29,10 +29,6 @@ common --host_cxxopt="--std=c++23"

 common --javacopt="-Xlint:-options"

-# std::filesystem and other modern C++ deps require macOS 10.15+.
-common:macos --macos_minimum_os=10.15
-common:macos --host_macos_minimum_os=10.15
-
 # suppress warnings due to https://developer.apple.com/forums/thread/733317
 # Use host_linkopt for macOS-specific flags to avoid passing them to Linux cross-compilation
 common:macos --host_linkopt=-Wl,-no_warn_duplicate_libraries
@@ -41,17 +37,11 @@ common:macos --host_linkopt=-Wl,-no_warn_duplicate_libraries
 # See: https://github.com/grpc/grpc/issues/37619
 common:macos --features=-module_maps

-# Pin DEVELOPER_DIR so Apple repo rules use full Xcode instead of Command Line
-# Tools. The sync script writes .bazelrc.xcode with an Xcode build-version
-# marker so in-place Xcode updates refresh Bazel's generated Apple toolchains.
+# Pin DEVELOPER_DIR so the apple_cc_autoconf repo rule doesn't re-evaluate
+# when Xcode updates in-place. The sync script overrides this for mactools.
 common:macos --repo_env=DEVELOPER_DIR=/Applications/Xcode.app/Contents/Developer

-# Local machine or CI overrides. This lets Bazel-only runners use Command Line
-# Tools without installing full Xcode. This must stay before .bazelrc.xcode so
-# generated Xcode settings win when both files exist.
-try-import %workspace%/.bazelrc.local
-
-# Xcode config for macOS builds. Generated by scripts/sync_bazel_xcode.sh.
+# Xcode config for mactools builds only. Generated by scripts/sync_bazel_xcode.sh.
 # Bakes the Xcode build version into action cache keys and sets DEVELOPER_DIR.
 try-import %workspace%/.bazelrc.xcode

@@ -6,32 +6,28 @@ on:
    - cron: '0 */6 * * *'
  workflow_dispatch:

-permissions:
-  contents: read
-  actions: write
-
 jobs:
  cleanup-expired:
    runs-on: ubuntu-latest

    steps:
-      - name: Delete expired artifacts and artifacts older than 3 days
+      - name: Delete artifacts older than 3 days
        env:
          GH_TOKEN: ${{ github.token }}
        run: |
          echo "Fetching all artifacts..."
          gh api "repos/${{ github.repository }}/actions/artifacts" \
-            --paginate -q '.artifacts[] | "\(.id)\t\(.created_at)\t\(.expired)\t\(.name)"' > /tmp/all_artifacts.txt
+            --paginate -q '.artifacts[] | "\(.id)\t\(.created_at)\t\(.name)"' > /tmp/all_artifacts.txt

          total=$(wc -l < /tmp/all_artifacts.txt)
          echo "Found $total total artifacts"

          cutoff=$(date -u -d '3 days ago' '+%Y-%m-%dT%H:%M:%SZ')
-          echo "Deleting expired artifacts and artifacts created before $cutoff"
+          echo "Deleting artifacts created before $cutoff"

          deleted=0
-          while IFS=$'\t' read -r id created_at expired name; do
-            if [[ "$expired" == "true" || "$created_at" < "$cutoff" ]]; then
+          while IFS=$'\t' read -r id created_at name; do
+            if [[ "$created_at" < "$cutoff" ]]; then
              gh api -X DELETE "repos/${{ github.repository }}/actions/artifacts/$id" 2>/dev/null && deleted=$((deleted + 1))
              if [ $((deleted % 50)) -eq 0 ]; then
                echo "Deleted $deleted artifacts so far..."
@@ -39,7 +35,7 @@ jobs:
            fi
          done < /tmp/all_artifacts.txt

-          echo "Cleanup complete. Deleted $deleted artifacts out of $total total."
+          echo "Cleanup complete. Deleted $deleted expired artifacts out of $total total."
          rm -f /tmp/all_artifacts.txt

  check-storage:
@@ -51,12 +47,11 @@ jobs:
        env:
          GH_TOKEN: ${{ github.token }}
        run: |
-          # Calculate active artifact storage. The artifacts API can list expired
-          # artifacts until they are explicitly deleted, so do not count them.
+          # Calculate total artifact storage
          total_bytes=$(gh api "repos/${{ github.repository }}/actions/artifacts" \
-            --paginate -q '.artifacts[] | select(.expired == false) | .size_in_bytes' | awk '{sum+=$1} END {print sum}')
+            --paginate -q '.artifacts[].size_in_bytes' | awk '{sum+=$1} END {print sum}')

-          total_mb=$(( ${total_bytes:-0} / 1024 / 1024 ))
+          total_mb=$((total_bytes / 1024 / 1024))
          echo "Total artifact storage: ${total_mb} MB"

          # Fail if over 500MB
@@ -66,7 +61,7 @@ jobs:
            echo "Largest artifacts:"
            # Save to temp file to avoid SIGPIPE/broken pipe errors with head
            gh api "repos/${{ github.repository }}/actions/artifacts" \
-              --paginate -q '.artifacts[] | select(.expired == false) | "\(.size_in_bytes)\t\(.name)\t\(.created_at)"' > /tmp/artifacts.txt
+              --paginate -q '.artifacts[] | "\(.size_in_bytes)\t\(.name)\t\(.created_at)"' > /tmp/artifacts.txt
            sort -rn /tmp/artifacts.txt | head -20 | \
              awk -F'\t' '{printf "%d MB\t%s\t%s\n", $1/1024/1024, $2, $3}'
            rm -f /tmp/artifacts.txt
@@ -11,7 +11,6 @@ on:
      - 'src/main/protobuf/net/eagle0/eagle/internal/auth_internal.proto'
      - 'src/main/resources/net/eagle0/attributions.json'
      - 'ci/BUILD.bazel'
-      - 'ci/github_actions/ensure_bazel_installed.sh'
      - '.bazelrc'
      - '.github/workflows/auth_build.yml'
  workflow_dispatch:
@@ -25,39 +24,17 @@ on:
 permissions:
  contents: read

-concurrency:
-  group: auth-build-deploy
-  cancel-in-progress: false
-
 jobs:
  build-auth:
    runs-on: [self-hosted, bazel]
    outputs:
      image_tag: ${{ steps.push-auth.outputs.image_tag }}
    steps:
-      - name: Prepare non-LFS checkout
-        env:
-          GIT_CONFIG_GLOBAL: ${{ runner.temp }}/gitconfig-no-lfs
-          GIT_CONFIG_NOSYSTEM: "1"
-        run: |
-          git config --global --unset-all filter.lfs.process || true
-          git config --global filter.lfs.smudge "cat"
-          git config --global filter.lfs.clean "cat"
-          git config --global filter.lfs.required false
-          rm -f .git/hooks/post-checkout .git/hooks/post-merge .git/hooks/pre-push
-
      - name: Checkout repository
        uses: actions/checkout@v6
-        env:
-          GIT_CONFIG_GLOBAL: ${{ runner.temp }}/gitconfig-no-lfs
-          GIT_CONFIG_NOSYSTEM: "1"
        with:
-          persist-credentials: false
          lfs: false

-      - name: Ensure Bazel installed
-        run: ./ci/github_actions/ensure_bazel_installed.sh
-
      - name: Build Auth Server Docker image
        id: build-auth
        run: |
@@ -151,8 +128,6 @@ jobs:
    steps:
      - name: Checkout repository
        uses: actions/checkout@v6
-        with:
-          persist-credentials: false

      - name: Deploy auth service to production
        uses: appleboy/ssh-action@v1.2.5
@@ -163,7 +138,7 @@ jobs:
          script_stop: true
          envs: AUTH_IMAGE,DISCORD_CLIENT_ID,DISCORD_CLIENT_SECRET,GOOGLE_CLIENT_ID,GOOGLE_CLIENT_SECRET,GH_OAUTH_CLIENT_ID,GH_OAUTH_CLIENT_SECRET,APPLE_SIGNIN_CLIENT_ID,APPLE_TEAM_ID,APPLE_SIGNIN_KEY_ID,APPLE_SIGNIN_PRIVATE_KEY,TWITCH_CLIENT_ID,TWITCH_CLIENT_SECRET,JWT_PRIVATE_KEY,FASTMAIL_API_TOKEN,FASTMAIL_FROM_EMAIL,FASTMAIL_FROM_NAME
          script: |
-            set -e
+            set -x
            cd /opt/eagle0

            # Export env vars for docker compose (appleboy/ssh-action sets them but doesn't export)
@@ -16,7 +16,6 @@ jobs:
      - name: Checkout
        uses: actions/checkout@v6
        with:
-          persist-credentials: false
          lfs: false

      - name: Show disk usage before cleanup
@@ -9,7 +9,6 @@ on:
      - 'MODULE.bazel'
      - 'BUILD.bazel'
      - '.bazelrc'
-      - 'ci/github_actions/ensure_bazel_installed.sh'
      - '.github/workflows/bazel_test.yml'
      - '!src/main/csharp/**'
      - '!src/test/csharp/**'
@@ -20,7 +19,6 @@ on:
      - 'MODULE.bazel'
      - 'BUILD.bazel'
      - '.bazelrc'
-      - 'ci/github_actions/ensure_bazel_installed.sh'
      - '.github/workflows/bazel_test.yml'
      - '!src/main/csharp/**'
      - '!src/test/csharp/**'
@@ -42,27 +40,9 @@ jobs:
        git config --local core.sparseCheckout false 2>/dev/null || true
        git config --local --unset extensions.worktreeConfig 2>/dev/null || true
        rm -f .git/info/sparse-checkout .git/config.worktree 2>/dev/null || true
-    - name: Ensure Git LFS available for checkout
-      run: |
-        COMMON_PATHS=(/opt/homebrew/bin /usr/local/bin)
-        for path in "${COMMON_PATHS[@]}"; do
-          if [ -d "$path" ]; then
-            echo "$path" >> "$GITHUB_PATH"
-            export PATH="$path:$PATH"
-          fi
-        done
-
-        if command -v git-lfs >/dev/null 2>&1; then
-          git-lfs version
-          exit 0
-        fi
-
-        brew install git-lfs
-        git-lfs version
    - name: Checkout repository
      uses: actions/checkout@v6
      with:
-        persist-credentials: false
        lfs: false
    - name: Check BUILD.bazel dependencies
      run: ./scripts/check_build_deps.sh --strict
@@ -82,30 +62,10 @@ jobs:
        git config --local core.sparseCheckout false 2>/dev/null || true
        git config --local --unset extensions.worktreeConfig 2>/dev/null || true
        rm -f .git/info/sparse-checkout .git/config.worktree 2>/dev/null || true
-    - name: Ensure Git LFS available for checkout
-      run: |
-        COMMON_PATHS=(/opt/homebrew/bin /usr/local/bin)
-        for path in "${COMMON_PATHS[@]}"; do
-          if [ -d "$path" ]; then
-            echo "$path" >> "$GITHUB_PATH"
-            export PATH="$path:$PATH"
-          fi
-        done
-
-        if command -v git-lfs >/dev/null 2>&1; then
-          git-lfs version
-          exit 0
-        fi
-
-        brew install git-lfs
-        git-lfs version
    - name: Checkout repository
      uses: actions/checkout@v6
      with:
-        persist-credentials: false
        lfs: false
-    - name: Ensure Bazel installed
-      run: ./ci/github_actions/ensure_bazel_installed.sh
    - name: Run tests
      run: bazel test --build_event_json_file=test.json //src/test/... //src/main/go/...
    - name: Collect failed test logs
@@ -16,13 +16,9 @@ jobs:
    steps:
      - uses: actions/checkout@v6
        with:
-          persist-credentials: false
          lfs: false
          clean: false

-      - name: Ensure Bazel installed
-        run: ./ci/github_actions/ensure_bazel_installed.sh
-
      - name: Clean up unreferenced blobs
        env:
          ACCESS_KEY_ID: ${{ secrets.ACCESS_KEY_ID }}
@@ -28,8 +28,6 @@ jobs:
    steps:
      - name: Checkout repository
        uses: actions/checkout@v6
-        with:
-          persist-credentials: false

      - name: Build sysroot
        run: ./tools/sysroot/build_sysroot.sh
@@ -84,8 +82,6 @@ jobs:
    steps:
      - name: Checkout repository
        uses: actions/checkout@v6
-        with:
-          persist-credentials: false

      - name: Set up QEMU for ARM64 emulation
        uses: docker/setup-qemu-action@v4
@@ -18,7 +18,6 @@ on:
      - '!src/main/protobuf/net/eagle0/eagle/internal/auth_internal.proto'
      - 'src/main/resources/**'
      - 'ci/BUILD.bazel'
-      - 'ci/github_actions/ensure_bazel_installed.sh'
      - 'MODULE.bazel'
      - '.bazelrc'
      - 'docker-compose.prod.yml'
@@ -69,38 +68,14 @@ jobs:
        env:
          GH_TOKEN: ${{ github.token }}

-      - name: Ensure Git LFS available for checkout
-        if: steps.check-latest.outputs.skip != 'true'
-        run: |
-          COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-          export PATH="${COMMON_PATHS}:${PATH}"
-          for path in /opt/homebrew/bin /usr/local/bin; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-            fi
-          done
-          if ! command -v git-lfs >/dev/null 2>&1; then
-            brew install git-lfs
-          fi
-          git-lfs --version
-
      - name: Checkout repository
        if: steps.check-latest.outputs.skip != 'true'
        uses: actions/checkout@v6
-        env:
-          GIT_LFS_SKIP_SMUDGE: 1
        with:
-          persist-credentials: false
          lfs: false

-      - name: Ensure Bazel installed
-        if: steps.check-latest.outputs.skip != 'true'
-        run: ./ci/github_actions/ensure_bazel_installed.sh
-
      - name: Fetch LFS files needed for admin server
        if: steps.check-latest.outputs.skip != 'true'
-        env:
-          GITHUB_TOKEN: ${{ github.token }}
        run: ./ci/github_actions/fetch_lfs.sh --include="src/main/go/net/eagle0/admin_server/static/tiles/*"

      - name: Build all Docker images
@@ -264,37 +239,14 @@ jobs:
      FASTMAIL_API_TOKEN: ${{ secrets.FASTMAIL_API_TOKEN }}
      FASTMAIL_FROM_EMAIL: ${{ secrets.FASTMAIL_FROM_EMAIL }}
      FASTMAIL_FROM_NAME: ${{ secrets.FASTMAIL_FROM_NAME }}
-      EAGLE_HISTORY_BACKEND: ${{ secrets.EAGLE_HISTORY_BACKEND }}
-      EAGLE_POSTGRES_HOST: ${{ secrets.EAGLE_POSTGRES_HOST }}
-      EAGLE_POSTGRES_PORT: ${{ secrets.EAGLE_POSTGRES_PORT }}
-      EAGLE_POSTGRES_DATABASE: ${{ secrets.EAGLE_POSTGRES_DATABASE }}
-      EAGLE_POSTGRES_USER: ${{ secrets.EAGLE_POSTGRES_USER }}
-      EAGLE_POSTGRES_PASSWORD: ${{ secrets.EAGLE_POSTGRES_PASSWORD }}
      DO_DROPLET_IP: ${{ secrets.DO_DROPLET_IP }}
      DO_REGISTRY_TOKEN: ${{ secrets.DO_REGISTRY_TOKEN }}
      NOTIFY_SECRET: ${{ secrets.EAGLE_NOTIFY_SECRET }}
      GITHUB_TOKEN_FOR_ADMIN: ${{ secrets.ADMIN_GITHUB_TOKEN }}
    steps:
-      - name: Ensure Git LFS available for checkout
-        run: |
-          COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-          export PATH="${COMMON_PATHS}:${PATH}"
-          for path in /opt/homebrew/bin /usr/local/bin; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-            fi
-          done
-          if ! command -v git-lfs >/dev/null 2>&1; then
-            brew install git-lfs
-          fi
-          git-lfs --version
-
      - name: Checkout repository
        uses: actions/checkout@v6
-        env:
-          GIT_LFS_SKIP_SMUDGE: 1
        with:
-          persist-credentials: false
          lfs: false

      - name: Setup SSH key
@@ -305,25 +257,11 @@ jobs:
          ssh-keyscan -H "$DO_DROPLET_IP" >> ~/.ssh/known_hosts 2>/dev/null || true

      - name: Download warmup binary
-        id: download-warmup
-        continue-on-error: true
        uses: actions/download-artifact@v8
        with:
          name: warmup-binary
          path: scripts/bin/

-      - name: Build warmup binary fallback
-        if: steps.download-warmup.outcome != 'success'
-        run: |
-          echo "::warning::warmup-binary artifact was unavailable; rebuilding warmup binary before deploy"
-          ./ci/github_actions/ensure_bazel_installed.sh
-          bazel build \
-            --platforms=//:linux_x86_64 \
-            --extra_toolchains=@llvm_toolchain_linux//:all \
-            //ci:warmup_tar
-          mkdir -p scripts/bin
-          tar -xf bazel-bin/ci/warmup_tar.tar -C scripts/bin --strip-components=1
-
      - name: Copy config files to droplet
        run: |
          # Create directory structure on remote
@@ -342,7 +280,7 @@ jobs:
      - name: Deploy to production droplet
        run: |
          ssh -i ~/.ssh/deploy_key deploy@"$DO_DROPLET_IP" bash -s << DEPLOY_SCRIPT
-          set -e
+          set -ex
          cd /opt/eagle0

          # =================================================================
@@ -380,10 +318,6 @@ jobs:
          validate_env "EAGLE_IMAGE" "${EAGLE_IMAGE}" "" || VALIDATION_FAILED=1
          validate_env "JWT_PRIVATE_KEY" "${JWT_PRIVATE_KEY}" "" || VALIDATION_FAILED=1
          validate_env "DO_REGISTRY_TOKEN" "${DO_REGISTRY_TOKEN}" "" || VALIDATION_FAILED=1
-          validate_env "EAGLE_POSTGRES_HOST" "${EAGLE_POSTGRES_HOST}" "" || VALIDATION_FAILED=1
-          validate_env "EAGLE_POSTGRES_DATABASE" "${EAGLE_POSTGRES_DATABASE}" "" || VALIDATION_FAILED=1
-          validate_env "EAGLE_POSTGRES_USER" "${EAGLE_POSTGRES_USER}" "" || VALIDATION_FAILED=1
-          validate_env "EAGLE_POSTGRES_PASSWORD" "${EAGLE_POSTGRES_PASSWORD}" "" || VALIDATION_FAILED=1

          if [ "\${VALIDATION_FAILED}" -eq 1 ]; then
              echo ""
@@ -431,12 +365,6 @@ jobs:
          export FASTMAIL_API_TOKEN="${FASTMAIL_API_TOKEN}"
          export FASTMAIL_FROM_EMAIL="${FASTMAIL_FROM_EMAIL}"
          export FASTMAIL_FROM_NAME="${FASTMAIL_FROM_NAME}"
-          export EAGLE_HISTORY_BACKEND="${EAGLE_HISTORY_BACKEND:-postgres}"
-          export EAGLE_POSTGRES_HOST="${EAGLE_POSTGRES_HOST}"
-          export EAGLE_POSTGRES_PORT="${EAGLE_POSTGRES_PORT}"
-          export EAGLE_POSTGRES_DATABASE="${EAGLE_POSTGRES_DATABASE}"
-          export EAGLE_POSTGRES_USER="${EAGLE_POSTGRES_USER}"
-          export EAGLE_POSTGRES_PASSWORD="${EAGLE_POSTGRES_PASSWORD}"
          export DO_REGISTRY_TOKEN="${DO_REGISTRY_TOKEN}"
          export NOTIFY_SECRET="${NOTIFY_SECRET}"
          export GITHUB_TOKEN="${GITHUB_TOKEN_FOR_ADMIN}"
@@ -11,7 +11,6 @@ on:
      - 'MODULE.bazel'
      - 'BUILD.bazel'
      - '.bazelrc'
-      - 'ci/github_actions/ensure_bazel_installed.sh'
      - '.github/workflows/eagle_build.yml'

 concurrency:
@@ -26,29 +25,9 @@ jobs:
    runs-on: [self-hosted, bazel]

    steps:
-      - name: Ensure Git LFS available for checkout
-        run: |
-          COMMON_PATHS=(/opt/homebrew/bin /usr/local/bin)
-          for path in "${COMMON_PATHS[@]}"; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-              export PATH="$path:$PATH"
-            fi
-          done
-
-          if command -v git-lfs >/dev/null 2>&1; then
-            git-lfs version
-            exit 0
-          fi
-
-          brew install git-lfs
-          git-lfs version
      - name: Checkout repository
        uses: actions/checkout@v6
        with:
-          persist-credentials: false
          lfs: false
-      - name: Ensure Bazel installed
-        run: ./ci/github_actions/ensure_bazel_installed.sh
      - name: Build Eagle server
        run: bazel build //src/main/scala/net/eagle0/eagle:eagle_server
@@ -5,12 +5,10 @@ on:
    branches: [ "main" ]
    paths:
      - ".github/workflows/installer_build.yml"
-      - "ci/github_actions/ensure_bazel_installed.sh"
      - "src/main/go/net/eagle0/clients/win/installer/**"
  pull_request:
    paths:
      - ".github/workflows/installer_build.yml"
-      - "ci/github_actions/ensure_bazel_installed.sh"
      - "src/main/go/net/eagle0/clients/win/installer/**"
  workflow_dispatch:

@@ -23,32 +21,11 @@ jobs:
    runs-on: [self-hosted, bazel]

    steps:
-      - name: Ensure Git LFS available for checkout
-        run: |
-          COMMON_PATHS=(/opt/homebrew/bin /usr/local/bin)
-          for path in "${COMMON_PATHS[@]}"; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-              export PATH="$path:$PATH"
-            fi
-          done
-
-          if command -v git-lfs >/dev/null 2>&1; then
-            git-lfs version
-            exit 0
-          fi
-
-          brew install git-lfs
-          git-lfs version
      - uses: actions/checkout@v6
        with:
-          persist-credentials: false
          lfs: false
          clean: false

-      - name: Ensure Bazel installed
-        run: ./ci/github_actions/ensure_bazel_installed.sh
-
      - name: Build Go installer for Windows
        env:
          MANIFEST_PUBLIC_KEY: ${{ secrets.MANIFEST_PUBLIC_KEY }}
@@ -13,7 +13,7 @@ on:

 permissions:
  contents: read
-  actions: write
+  actions: read

 env:
  # Runner-specific build directory to allow parallel builds on multiple runners
@@ -122,26 +122,8 @@ jobs:
            git for-each-ref --format='%(refname)' refs/remotes/pull/ 2>/dev/null | \
              xargs -r git update-ref -d 2>/dev/null || true
          fi
-
-      - name: Ensure Git LFS available for checkout
-        run: |
-          COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-          export PATH="${COMMON_PATHS}:${PATH}"
-          for path in /opt/homebrew/bin /usr/local/bin; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-            fi
-          done
-          if ! command -v git-lfs >/dev/null 2>&1; then
-            brew install git-lfs
-          fi
-          git-lfs --version
-
      - uses: actions/checkout@v6
-        env:
-          GIT_LFS_SKIP_SMUDGE: 1
        with:
-          persist-credentials: false
          lfs: false  # Fetch LFS after checkout to avoid stale ref issues
          clean: false  # Library/ persists between runs on self-hosted runners
          fetch-depth: 0
@@ -152,8 +134,6 @@ jobs:
          rm -rf src/main/csharp/net/eagle0/clients/unity/eagle0/Library/Bee/

      - name: Fetch LFS files
-        env:
-          GITHUB_TOKEN: ${{ github.token }}
        run: ./ci/github_actions/fetch_lfs.sh

      - name: Ensure Unity version installed
@@ -190,75 +170,11 @@ jobs:
          path: ${{ env.EAGLE0_BUILD_DIR }}/editor_ios.log
          retention-days: 3

-      - name: Package generated iOS project
-        if: success()
-        run: |
-          tar -czf "$RUNNER_TEMP/eagle0-ios-project.tar.gz" -C "$EAGLE0_BUILD_DIR" eagle0iOS
-
-      - name: Upload generated iOS project
-        if: success()
-        uses: actions/upload-artifact@v7
-        with:
-          name: eagle0-ios-project-${{ github.run_id }}
-          path: ${{ runner.temp }}/eagle0-ios-project.tar.gz
-          retention-days: 1
-
-      - name: Cleanup build directory
-        if: always()
-        run: rm -rf "${{ env.EAGLE0_BUILD_DIR }}"
-
  archive-and-upload:
    needs: build-unity
    runs-on: [self-hosted, macOS, testflight]

    steps:
-      - name: Ensure Git LFS available for checkout
-        run: |
-          COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-          export PATH="${COMMON_PATHS}:${PATH}"
-          for path in /opt/homebrew/bin /usr/local/bin; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-            fi
-          done
-          if ! command -v git-lfs >/dev/null 2>&1; then
-            brew install git-lfs
-          fi
-          git-lfs --version
-
-      - uses: actions/checkout@v6
-        env:
-          GIT_LFS_SKIP_SMUDGE: 1
-        with:
-          persist-credentials: false
-          lfs: false
-
-      - name: Clean build directory
-        run: rm -rf "${{ env.EAGLE0_BUILD_DIR }}"
-
-      - name: Download generated iOS project
-        uses: actions/download-artifact@v8
-        with:
-          name: eagle0-ios-project-${{ github.run_id }}
-          path: ${{ runner.temp }}
-
-      - name: Extract generated iOS project
-        run: |
-          mkdir -p "$EAGLE0_BUILD_DIR"
-          tar -xzf "$RUNNER_TEMP/eagle0-ios-project.tar.gz" -C "$EAGLE0_BUILD_DIR"
-
-      - name: Delete generated iOS project artifact
-        env:
-          GH_TOKEN: ${{ github.token }}
-        run: |
-          artifact_id=$(gh api "repos/${{ github.repository }}/actions/artifacts" \
-            --paginate \
-            --jq '.artifacts[] | select(.name == "eagle0-ios-project-${{ github.run_id }}") | .id')
-
-          if [ -n "$artifact_id" ]; then
-            gh api -X DELETE "repos/${{ github.repository }}/actions/artifacts/$artifact_id"
-          fi
-
      - name: Install Signing Certificate
        env:
          IOS_CERTIFICATE: ${{ secrets.IOS_CERTIFICATE }}
@@ -20,7 +20,6 @@ on:
      - "scripts/notarize_wait.sh"
      - "ci/github_actions/build_mac.sh"
      - "ci/github_actions/build_unity_mac.sh"
-      - "ci/github_actions/ensure_bazel_installed.sh"
      - "ci/github_actions/upload_addressables.sh"
      - "ci/github_actions/ensure_unity_installed.sh"
      - "src/main/csharp/net/eagle0/clients/unity/eagle0/ProjectSettings/ProjectVersion.txt"
@@ -40,7 +39,6 @@ on:
      - "scripts/notarize_wait.sh"
      - "ci/github_actions/build_mac.sh"
      - "ci/github_actions/build_unity_mac.sh"
-      - "ci/github_actions/ensure_bazel_installed.sh"
      - "ci/mac/**"
  workflow_dispatch:
    inputs:
@@ -83,26 +81,8 @@ jobs:
            git for-each-ref --format='%(refname)' refs/remotes/pull/ 2>/dev/null | \
              xargs -r git update-ref -d 2>/dev/null || true
          fi
-
-      - name: Ensure Git LFS available for checkout
-        run: |
-          COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-          export PATH="${COMMON_PATHS}:${PATH}"
-          for path in /opt/homebrew/bin /usr/local/bin; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-            fi
-          done
-          if ! command -v git-lfs >/dev/null 2>&1; then
-            brew install git-lfs
-          fi
-          git-lfs --version
-
      - uses: actions/checkout@v6
-        env:
-          GIT_LFS_SKIP_SMUDGE: 1
        with:
-          persist-credentials: false
          lfs: false  # Fetch LFS after checkout to avoid stale ref issues
          clean: false  # Library/ persists between runs on self-hosted runners
          fetch-depth: 0  # For version numbering from git history
@@ -129,16 +109,11 @@ jobs:
          fi

      - name: Fetch LFS files
-        env:
-          GITHUB_TOKEN: ${{ github.token }}
        run: ./ci/github_actions/fetch_lfs.sh

      - name: Ensure Unity version installed
        run: ./ci/github_actions/ensure_unity_installed.sh mac

-      - name: Ensure Bazel installed
-        run: ./ci/github_actions/ensure_bazel_installed.sh
-
      - name: Sync Bazel Xcode config
        run: ./scripts/sync_bazel_xcode.sh

@@ -147,9 +122,7 @@ jobs:

      - name: Save build SHA for Bee/ cache invalidation
        if: success()
-        run: |
-          mkdir -p src/main/csharp/net/eagle0/clients/unity/eagle0/Library
-          git rev-parse HEAD > src/main/csharp/net/eagle0/clients/unity/eagle0/Library/.last_built_sha
+        run: git rev-parse HEAD > src/main/csharp/net/eagle0/clients/unity/eagle0/Library/.last_built_sha

      - name: Upload Addressables to CDN
        if: success() && github.ref == 'refs/heads/main' && github.event_name != 'pull_request'
@@ -277,25 +250,8 @@ jobs:
    runs-on: [self-hosted, macOS, notarize]

    steps:
-      - name: Ensure Git LFS available for checkout
-        run: |
-          COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-          export PATH="${COMMON_PATHS}:${PATH}"
-          for path in /opt/homebrew/bin /usr/local/bin; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-            fi
-          done
-          if ! command -v git-lfs >/dev/null 2>&1; then
-            brew install git-lfs
-          fi
-          git-lfs --version
-
      - uses: actions/checkout@v6
-        env:
-          GIT_LFS_SKIP_SMUDGE: 1
        with:
-          persist-credentials: false
          sparse-checkout: scripts

      - name: Clean download directory
@@ -342,30 +298,13 @@ jobs:
  deploy:
    needs: [build-and-sign, wait-notarization]
    if: needs.build-and-sign.outputs.should_deploy == 'true'
-    runs-on: [self-hosted, macOS, notarize]
+    runs-on: [self-hosted, macOS, unity-mac]
    outputs:
      deployed_version: ${{ steps.deploy-mac.outputs.deployed_version }}

    steps:
-      - name: Ensure Git LFS available for checkout
-        run: |
-          COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-          export PATH="${COMMON_PATHS}:${PATH}"
-          for path in /opt/homebrew/bin /usr/local/bin; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-            fi
-          done
-          if ! command -v git-lfs >/dev/null 2>&1; then
-            brew install git-lfs
-          fi
-          git-lfs --version
-
      - uses: actions/checkout@v6
-        env:
-          GIT_LFS_SKIP_SMUDGE: 1
        with:
-          persist-credentials: false
          fetch-depth: 0  # For version numbering

      - name: Clean download directory
@@ -383,9 +322,6 @@ jobs:
          ditto -x -k eagle0.app.zip .
          rm eagle0.app.zip

-      - name: Ensure Bazel installed
-        run: ./ci/github_actions/ensure_bazel_installed.sh
-
      - name: Sync Bazel Xcode config
        run: ./scripts/sync_bazel_xcode.sh

@@ -401,13 +337,8 @@ jobs:
          echo "$SPARKLE_EDDSA_PRIVATE_KEY" > "$SPARKLE_PRIVATE_KEY_PATH"
          chmod 600 "$SPARKLE_PRIVATE_KEY_PATH"

-          # Install dmgbuild in an isolated venv. Homebrew Python rejects system
-          # installs under PEP 668, and mac_build_handler shells out to dmgbuild.
-          DMGBUILD_VENV="${{ env.EAGLE0_BUILD_DIR }}/dmgbuild-venv"
-          python3 -m venv "$DMGBUILD_VENV"
-          "$DMGBUILD_VENV/bin/python" -m pip install --upgrade pip
-          "$DMGBUILD_VENV/bin/python" -m pip install dmgbuild
-          export PATH="$DMGBUILD_VENV/bin:$PATH"
+          # Install dmgbuild (creates .DS_Store programmatically, no AppleScript needed)
+          pip3 install dmgbuild

          # Background image for styled DMG
          BACKGROUND_PATH="$(pwd)/ci/mac/dmg/background.png"
@@ -39,9 +39,9 @@ jobs:
          echo "Dry run: ${DRY_RUN}"
          echo ""

-          # List repositories via JSON. Text output can wrap table rows and make
-          # headers/tags/digests look like repository names.
-          REPOS=$(doctl registry repository list-v2 "${REGISTRY}" --output json | jq -r '.[] | .name // .Name // empty')
+          # List of repositories to clean
+          # Use tail to skip header row in case --no-header doesn't work
+          REPOS=$(doctl registry repository list-v2 --format Name --no-header | grep -v '^Name$' | grep -v '^$')

          for REPO in $REPOS; do
            echo "=== Processing repository: ${REPO} ==="
@@ -88,20 +88,5 @@ jobs:
        if: github.event_name == 'schedule' || (github.event_name == 'workflow_dispatch' && github.event.inputs.dry_run == 'false')
        run: |
          echo "Starting garbage collection..."
-          set +e
-          GC_OUTPUT=$(doctl registry garbage-collection start --force 2>&1)
-          GC_STATUS=$?
-          set -e
-
-          echo "$GC_OUTPUT"
-          if [ "$GC_STATUS" -eq 0 ]; then
-            echo "Garbage collection started. It may take a few minutes to complete."
-            exit 0
-          fi
-
-          if echo "$GC_OUTPUT" | grep -q "automated garbage collection is enabled"; then
-            echo "Automated garbage collection is enabled for this registry; skipping manual garbage collection."
-            exit 0
-          fi
-
-          exit "$GC_STATUS"
+          doctl registry garbage-collection start --force
+          echo "Garbage collection started. It may take a few minutes to complete."
@@ -9,7 +9,6 @@ on:
      - 'src/main/protobuf/net/eagle0/common/**'
      - 'src/main/resources/net/eagle0/shardok/**'
      - 'ci/BUILD.bazel'
-      - 'ci/github_actions/ensure_bazel_installed.sh'
      - 'MODULE.bazel'
      - '.bazelrc'
      - '.github/workflows/shardok_arm64_build.yml'
@@ -34,29 +33,11 @@ jobs:
    outputs:
      image_tag: ${{ steps.push-shardok.outputs.image_tag }}
    steps:
-      - name: Prepare non-LFS checkout
-        env:
-          GIT_CONFIG_GLOBAL: ${{ runner.temp }}/gitconfig-no-lfs
-          GIT_CONFIG_NOSYSTEM: "1"
-        run: |
-          git config --global --unset-all filter.lfs.process || true
-          git config --global filter.lfs.smudge "cat"
-          git config --global filter.lfs.clean "cat"
-          git config --global filter.lfs.required false
-          rm -f .git/hooks/post-checkout .git/hooks/post-merge .git/hooks/pre-push
-
      - name: Checkout repository
        uses: actions/checkout@v6
-        env:
-          GIT_CONFIG_GLOBAL: ${{ runner.temp }}/gitconfig-no-lfs
-          GIT_CONFIG_NOSYSTEM: "1"
        with:
-          persist-credentials: false
          lfs: false

-      - name: Ensure Bazel installed
-        run: ./ci/github_actions/ensure_bazel_installed.sh
-
      - name: Build Shardok ARM64 binary (cross-compile for Linux ARM64)
        run: |
          set -ex
@@ -199,55 +180,33 @@ jobs:
          # Add host key to known_hosts to avoid prompt
          ssh-keyscan -H ${{ secrets.HETZNER_IP }} >> ~/.ssh/known_hosts 2>/dev/null || true

-      - name: Pull Shardok image tarball for Hetzner
-        env:
-          DO_REGISTRY_TOKEN: ${{ secrets.DO_REGISTRY_TOKEN }}
-        run: |
-          set -ex
-
-          # Hetzner is IPv6-only, while Docker can be redirected to IPv4-only
-          # DigitalOcean registry/blob endpoints. Pull on the runner and copy
-          # the Docker-loadable tarball over SSH instead.
-          CRANE_VERSION="v0.20.2"
-          CRANE_DIR="${HOME}/.local/bin"
-          CRANE="${CRANE_DIR}/crane"
-          if [ ! -x "$CRANE" ] || ! "$CRANE" version 2>/dev/null | grep -q "0.20.2"; then
-            echo "Installing crane ${CRANE_VERSION}..."
-            mkdir -p "$CRANE_DIR"
-            curl -sL "https://github.com/google/go-containerregistry/releases/download/${CRANE_VERSION}/go-containerregistry_Darwin_arm64.tar.gz" | tar xzf - -C "$CRANE_DIR" crane
-            chmod +x "$CRANE"
-          else
-            echo "Using cached crane at $CRANE"
-          fi
-
-          AUTH=$(echo -n "${DO_REGISTRY_TOKEN}:${DO_REGISTRY_TOKEN}" | base64)
-          mkdir -p ~/.docker
-          echo "{\"auths\":{\"registry.digitalocean.com\":{\"auth\":\"${AUTH}\"}}}" > ~/.docker/config.json
-
-          IMAGE_TAR="${RUNNER_TEMP}/shardok-arm64-image.tar"
-          "$CRANE" pull "${SHARDOK_IMAGE}" "$IMAGE_TAR"
-          echo "SHARDOK_IMAGE_TAR=$IMAGE_TAR" >> "$GITHUB_ENV"
-
-      - name: Copy Shardok image tarball to Hetzner
-        run: |
-          set -ex
-          ssh -i ~/.ssh/hetzner_deploy -o StrictHostKeyChecking=accept-new deploy@${{ secrets.HETZNER_IP }} \
-            "cat > /tmp/shardok-arm64-image.tar" < "$SHARDOK_IMAGE_TAR"
-
      - name: Deploy to Hetzner
        run: |
          ssh -i ~/.ssh/hetzner_deploy -o StrictHostKeyChecking=accept-new deploy@${{ secrets.HETZNER_IP }} << 'ENDSSH'
          set -ex
          cd /opt/eagle0

-          SHARDOK_IMAGE="${{ needs.build-shardok-arm64.outputs.image_tag }}"
-          echo "Loading Shardok ARM64 image: $SHARDOK_IMAGE"
-          docker load -i /tmp/shardok-arm64-image.tar
-          rm -f /tmp/shardok-arm64-image.tar
-          docker image inspect "$SHARDOK_IMAGE" > /dev/null
-          echo "Loaded image ID: $(docker image inspect --format '{{.Id}}' "$SHARDOK_IMAGE")"
-          echo "Shardok binary SHA256:"
-          docker run --rm --entrypoint sha256sum "$SHARDOK_IMAGE" /app/shardok-server
+          # Login to registry
+          echo "${{ secrets.DO_REGISTRY_TOKEN }}" | docker login registry.digitalocean.com -u "${{ secrets.DO_REGISTRY_TOKEN }}" --password-stdin
+
+          echo "Deploying Shardok ARM64: ${{ needs.build-shardok-arm64.outputs.image_tag }}"
+
+          # Pre-warm NAT64 path to DigitalOcean registry (IPv6-only Hetzner → IPv4 registry)
+          curl -sf --max-time 10 https://registry.digitalocean.com/v2/ > /dev/null 2>&1 || true
+
+          # Pull the new image (retry up to 3 times for NAT64 connectivity flakiness)
+          for attempt in 1 2 3; do
+            echo "Pull attempt $attempt..."
+            if docker pull "${{ needs.build-shardok-arm64.outputs.image_tag }}"; then
+              break
+            fi
+            if [ "$attempt" -eq 3 ]; then
+              echo "ERROR: docker pull failed after 3 attempts"
+              exit 1
+            fi
+            echo "Pull failed, retrying in 10s..."
+            sleep 10
+          done

          # Stop and remove any container using port 40042 or named shardok*
          docker ps -q --filter "publish=40042" | xargs -r docker stop
@@ -265,12 +224,11 @@ jobs:
            -v /usr/local/share/eagle0:/usr/local/share/eagle0:ro \
            -e SHARDOK_RESOURCES_PATH=/app/resources \
            -e SHARDOK_MAPS_PATH=/app/resources/maps \
-            "$SHARDOK_IMAGE"
+            "${{ needs.build-shardok-arm64.outputs.image_tag }}"

          # Wait and verify
          sleep 5
          docker ps | grep shardok-ai
-          echo "Running container image ID: $(docker inspect --format '{{.Image}}' shardok-ai)"

          # Cleanup old images
          docker image prune -f
@@ -5,14 +5,13 @@ on:
  pull_request:
    paths:
      - 'src/main/cpp/**'
-      - 'src/main/protobuf/net/eagle0/shardok/**'
-      - 'src/main/protobuf/net/eagle0/common/**'
+      - 'src/main/proto/net/eagle0/shardok/**'
+      - 'src/main/proto/net/eagle0/common/**'
      - 'src/main/go/net/eagle0/build/**'
      - 'WORKSPACE'
      - 'MODULE.bazel'
      - 'BUILD.bazel'
      - '.bazelrc'
-      - 'ci/github_actions/ensure_bazel_installed.sh'
      - '.github/workflows/shardok_build.yml'

 concurrency:
@@ -27,29 +26,9 @@ jobs:
    runs-on: [self-hosted, bazel]

    steps:
-      - name: Ensure Git LFS available for checkout
-        run: |
-          COMMON_PATHS=(/opt/homebrew/bin /usr/local/bin)
-          for path in "${COMMON_PATHS[@]}"; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-              export PATH="$path:$PATH"
-            fi
-          done
-
-          if command -v git-lfs >/dev/null 2>&1; then
-            git-lfs version
-            exit 0
-          fi
-
-          brew install git-lfs
-          git-lfs version
      - name: Checkout repository
        uses: actions/checkout@v6
        with:
-          persist-credentials: false
          lfs: false
-      - name: Ensure Bazel installed
-        run: ./ci/github_actions/ensure_bazel_installed.sh
      - name: Build Shardok server
        run: bazel build -c opt //src/main/cpp/net/eagle0/shardok:shardok-server
@@ -15,8 +15,6 @@ on:
      - "scripts/build_plugins.sh"
      - "scripts/build_windows_plugin.sh"
      - "ci/github_actions/build_unity.sh"
-      - "ci/github_actions/test_unity_editmode.sh"
-      - "ci/github_actions/ensure_bazel_installed.sh"
      - "ci/github_actions/upload_addressables.sh"
      - "ci/github_actions/ensure_unity_installed.sh"
      - "src/main/csharp/net/eagle0/clients/unity/eagle0/ProjectSettings/ProjectVersion.txt"
@@ -36,8 +34,6 @@ on:
      - "scripts/build_plugins.sh"
      - "scripts/build_windows_plugin.sh"
      - "ci/github_actions/build_unity.sh"
-      - "ci/github_actions/test_unity_editmode.sh"
-      - "ci/github_actions/ensure_bazel_installed.sh"
      - "ci/github_actions/upload_addressables.sh"
      - "ci/github_actions/ensure_unity_installed.sh"
      - "src/main/csharp/net/eagle0/clients/unity/eagle0/ProjectSettings/ProjectVersion.txt"
@@ -73,26 +69,8 @@ jobs:
            git for-each-ref --format='%(refname)' refs/remotes/pull/ 2>/dev/null | \
              xargs -r git update-ref -d 2>/dev/null || true
          fi
-
-      - name: Ensure Git LFS available for checkout
-        run: |
-          COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-          export PATH="${COMMON_PATHS}:${PATH}"
-          for path in /opt/homebrew/bin /usr/local/bin; do
-            if [ -d "$path" ]; then
-              echo "$path" >> "$GITHUB_PATH"
-            fi
-          done
-          if ! command -v git-lfs >/dev/null 2>&1; then
-            brew install git-lfs
-          fi
-          git-lfs --version
-
      - uses: actions/checkout@v6
-        env:
-          GIT_LFS_SKIP_SMUDGE: 1
        with:
-          persist-credentials: false
          lfs: false  # Fetch LFS after checkout to avoid stale ref issues
          clean: false  # Library/ persists between runs on self-hosted runners

@@ -132,28 +110,19 @@ jobs:
          fi

      - name: Fetch LFS files
-        env:
-          GITHUB_TOKEN: ${{ github.token }}
        run: ./ci/github_actions/fetch_lfs.sh
      - name: Ensure Unity version installed
        run: ./ci/github_actions/ensure_unity_installed.sh windows
-      - name: Ensure Bazel installed
-        run: ./ci/github_actions/ensure_bazel_installed.sh
-      - name: Run Unity EditMode tests
-        run: ./ci/github_actions/test_unity_editmode.sh
      - name: Build Windows unity
        run: ./ci/github_actions/build_unity.sh "${{ env.EAGLE0_BUILD_DIR }}/eagle0WIN"

      - name: Save build SHA for Bee/ cache invalidation
        if: success()
-        run: |
-          mkdir -p src/main/csharp/net/eagle0/clients/unity/eagle0/Library
-          git rev-parse HEAD > src/main/csharp/net/eagle0/clients/unity/eagle0/Library/.last_built_sha
+        run: git rev-parse HEAD > src/main/csharp/net/eagle0/clients/unity/eagle0/Library/.last_built_sha

      - name: Save Unity version for Library/ cache invalidation
        if: success()
        run: |
-          mkdir -p src/main/csharp/net/eagle0/clients/unity/eagle0/Library
          grep "m_EditorVersion:" src/main/csharp/net/eagle0/clients/unity/eagle0/ProjectSettings/ProjectVersion.txt | head -1 | sed 's/m_EditorVersion: //' > src/main/csharp/net/eagle0/clients/unity/eagle0/Library/.last_unity_version

      - name: Upload Addressables to CDN
@@ -207,15 +176,6 @@ jobs:
          name: editor_win.log
          path: ${{ env.EAGLE0_BUILD_DIR }}/editor_win.log
          retention-days: 3
-      - name: Archive EditMode test artifacts
-        if: success() || failure()
-        uses: actions/upload-artifact@v7
-        with:
-          name: unity_editmode_tests
-          path: |
-            ${{ env.EAGLE0_BUILD_DIR }}/editor_editmode_tests.log
-            ${{ env.EAGLE0_BUILD_DIR }}/editmode-test-results.xml
-          retention-days: 3
      - name: Cleanup build directory
        if: always()
        run: rm -rf "${{ env.EAGLE0_BUILD_DIR }}"
@@ -23,7 +23,6 @@ bazel-bin
 bazel-eagle0*
 bazel-out
 bazel-testlogs
-.bazelrc.local
 .bazelrc.xcode
 .ijwb
 .clwb
@@ -44,5 +43,3 @@ src/main/csharp/net/eagle0/clients/unity/eagle0/Assets/Plugins/Eagle0Protos/
 src/main/csharp/net/eagle0/clients/unity/eagle0/ProjectSettings/Packages/com.unity.dedicated-server/
 node_modules/
 tools/map_generator/output/
-src/main/csharp/net/eagle0/clients/unity/eagle0/docs/generated/unity_asset_usage_audit_files.csv
-src/main/csharp/net/eagle0/clients/unity/eagle0/docs/generated/unity_lfs_asset_audit.csv
@@ -34,8 +34,8 @@ repos:
    hooks:
    -   id: scalafmt
        name: scalafmt
-        language: system
-        entry: ./scripts/pre-commit-scalafmt.sh
+        language: system 
+        entry: scalafmt -i -f 
        types_or: ["scala"]
 - repo: local
  hooks:
@@ -1,391 +0,0 @@
-# AGENTS.md
-
-## CRITICAL BASH RULES (NEVER VIOLATE)
-
-**NEVER chain bash commands.** Do not use `&&`, `||`, or `;` to combine commands. Each command must be a separate Bash tool call. Use parallel tool calls when commands are independent.
-
-**NEVER prefix a command with `cd`.** Run `git`, `gh`, `bazel`, etc. directly from whatever the cwd already is. The cwd is a subdirectory of the worktree; `git`/`gh` find the repo via `.git` discovery and `bazel` finds the workspace via MODULE.bazel discovery — walking up from a subdir works fine. The shell resets cwd after every command anyway, so a `cd` never persists. (And `cd <dir> && <cmd>` violates the no-chaining rule above and forces an approval prompt every time.) Only `cd` in the rare case a tool genuinely cannot locate its root, and even then as its own Bash call, never a chain.
-
-## CRITICAL GIT RULES (NEVER VIOLATE)
-
-**NEVER use `git -C`.** Just run `git` directly from the cwd — it finds the repo via `.git` discovery. Do not `cd` to the repo root either (see the no-`cd` bash rule above).
-
-**NEVER push directly to main/master.** No exceptions. Not for "small changes." Not for docs. Not ever.
-
-**NEVER merge PRs.** You create PRs. The user merges them. No exceptions.
-
-**ALWAYS use this workflow:**
-1. Create a feature branch from origin/main
-2. Commit to that branch
-3. Create a PR with `gh pr create`
-4. Wait for user to merge (DO NOT run `gh pr merge`)
-
-If you catch yourself about to run `git push origin main` or `git push origin <branch>:main`, STOP. You are about to violate a critical rule. Create a PR instead.
-
-If you catch yourself about to run `gh pr merge`, STOP. Only the user merges PRs.
-
-## Worktree Usage
-
-Prefer using the current worktree for small requested fixes. Create a separate worktree only when the user asks for one,
-when the current branch/state would make the change unsafe, or when isolating a large/risky change is clearly beneficial.
-
-**NEVER leave the current worktree without explicit user approval.** Do not create, switch to, edit, commit in, or run
-git commands from another worktree unless the user has specifically asked for that worktree or approved the move. If the
-current worktree is unsafe because of dirty state, report that and ask before using a different worktree.
-
-Codex may edit code and run git commands in the current worktree, even when unrelated files are dirty, as long as it:
- does not revert, reset, clean, or overwrite unrelated changes
- stages only the files relevant to the requested change
- verifies the staged diff before committing
- creates feature branches and PRs as usual
- does not push to main/master or merge PRs
-
-When the user says a Codex-created PR has been merged, delete the corresponding local branch and temporary worktree
-automatically, unless there is uncommitted work that would make cleanup unsafe.
-
-## PR Timing
-
-When code changes appear correct locally, create the PR promptly even if long-running builds or tests are still running.
-Opening the PR starts CI on the build machines, uploads results to the Bazel remote cache, and can speed up subsequent
-local validation. Continue monitoring any already-started local and CI validation after opening the PR.
-
-## GitHub CLI PR Bodies
-
-When creating or editing PR descriptions with multiline bodies, write the body to a temporary markdown file and pass it
-with `gh pr create --body-file <path>` or `gh pr edit <number> --body-file <path>`. Do not use multiline `$'...'`
-shell quoting for PR bodies; Codex's permission matching may treat that as a complex shell command instead of a clean
-`gh pr` invocation, causing unnecessary approval prompts.
-
-Write these temporary body files, and any other scratch files needed for commands, under a known writable location such
-as `/private/tmp`, `$TMPDIR`, or the current repository workspace. Do not use protected paths that require additional
-user permission just to create or edit temporary files.
-
---
-
-This file provides guidance to Codex (Codex.ai/code) when working with code in this repository.
-
-## Project Overview
-
-Eagle0 is a multi-language gaming system combining strategic turn-based gameplay (Eagle) with tactical hex-based
-combat (Shardok). The system integrates LLM-based narrative generation and supports both human and AI players.
-
-## Architecture
-
-**Three-Tier Game System:**
-
- **Unity Client (C#)**: Real-time strategy game client with integrated tactical combat UI
- **Eagle (Scala)**: Strategic layer managing turn-based gameplay, diplomacy, hero progression, and province control
- **Shardok (C++)**: Tactical layer handling real-time hex-based combat simulation with performance-critical battle
-  resolution
-
-**Communication Flow:**
-
-```
-Unity Client ↔ Eagle (gRPC streaming) ↔ Shardok (internal gRPC) 
-```
-
-**Key Entry Points:**
-
- `/src/main/csharp/net/eagle0/clients/unity/eagle0/` - Unity C# game client
- `/src/main/scala/net/eagle0/eagle/Main.scala` - Eagle strategic game server
- `/src/main/cpp/net/eagle0/shardok/shardok_server_main.cpp` - Shardok tactical server
-
-**Protocol Buffer Architecture:**
-
- Extensive use of protobuf for type-safe communication
- Separate packages: `api/` (client-facing), `internal/` (server state), `views/` (client projections)
- Event sourcing pattern with immutable action history
-
-## Essential Commands
-
-### Building
-
-```bash
-# Build Eagle server (Scala strategic layer)
-bazel build //src/main/scala/net/eagle0/eagle:eagle_server_deploy.jar
-
-# Build Shardok server (C++ tactical layer)
-bazel build -c opt //src/main/cpp/net/eagle0/shardok:shardok-server
-
-# Shardok server includes both AI algorithms
-bazel build //src/main/cpp/net/eagle0/shardok:shardok-server
-
-# Build Unity/C# client
-./scripts/build_protos.sh         # Protocol buffer generation for Unity
-./scripts/build_plugins.sh        # Native plugins for all platforms  
-./scripts/build_windows_plugin.sh # Windows-specific plugin build
-# Unity builds via CI: ci/github_actions/build_unity.sh
-```
-
-### Running Services
-
-```bash
-# Eagle server (port 40032)
-bazel run //src/main/scala/net/eagle0/eagle:eagle_server -- --eagle-grpc-port 40032
-# Or: ./scripts/eagle_run.sh
-
-# Shardok server  
-bazel run //src/main/cpp/net/eagle0/shardok:shardok-server --compilation_mode=opt
-# Or: ./scripts/shardok_run.sh
-```
-
-### Testing
-
-```bash
-# Run all tests
-bazel test //src/test/... //src/main/go/...
-
-# Component-specific tests
-bazel test //src/test/scala/...  # Scala Eagle tests
-bazel test //src/test/cpp/...    # C++ Shardok tests
-```
-
-### Code Generation
-
-```bash
-bazel run gazelle              # Update Go build files
-./scripts/updateActionResultTypes.sh   # Update protocol buffer mappings
-```
-
-### Pre-Commit Checklist
-
-**MANDATORY: Before running `git commit`, verify:**
-
-1. **If you modified any BUILD.bazel file:** Run `bazel run gazelle` and stage any changes it makes
-2. **If you modified C++ or C# files:** Run `clang-format -i` on the modified files
-3. **If you modified Scala files:** scalafmt will run automatically via pre-commit hook
-4. **ALWAYS run the full test suite for your language changes before pushing:**
-   - For Scala changes: `bazel test //src/test/scala/...`
-   - For C++ changes: `bazel test //src/test/cpp/...`
-   - This catches exhaustive pattern match errors and other compile-time failures that single-target builds miss
-
-The pre-commit hook runs gazelle but only checks if it succeeds - it does NOT verify the BUILD files are in canonical format. The `gazelle_test` will fail if deps are not alphabetically sorted. **Always run gazelle manually after BUILD file changes.**
-
-### Code Formatting
-
-```bash
-# ALWAYS run clang-format after making any C++ or C# code changes
-clang-format -i <modified_files>
-
-# Format all C++ files in a directory:
-find . -name "*.cpp" -o -name "*.hpp" | xargs clang-format -i
-
-# Format all C# files in a directory:
-find . -name "*.cs" | xargs clang-format -i
-```
-
-### Static Analysis
-
-```bash
-# Run clang-tidy static analysis on C++ files
-# Note: This may show some header include errors but will still analyze the main file
-bazel run @llvm_toolchain//:clang-tidy -- --checks='readability-*,bugprone-*,clang-analyzer-*' <file_path> -- -I/Users/dancrosby/CodingProjects/github/eagle0 -std=c++23
-
-# Example for AI files:
-bazel run @llvm_toolchain//:clang-tidy -- --checks='readability-*,bugprone-*,clang-analyzer-*' /Users/dancrosby/CodingProjects/github/eagle0/src/main/cpp/net/eagle0/shardok/ai/AIUnitScoreCalculator.cpp -- -I/Users/dancrosby/CodingProjects/github/eagle0 -std=c++23
-```
-
-## AI Algorithm Selection
-
-Eagle0 supports two AI algorithms for tactical combat decision-making:
-
-### Iterative Deepening AI (Default)
-
-The original minimax-based AI with sophisticated randomness handling:
-
- **Advantages**: Proven, sophisticated randomness evaluation, comprehensive lookahead
- **Use cases**: Production builds, scenarios requiring precise evaluation
- **Performance**: Single-threaded, thorough evaluation
-
-### Monte Carlo Tree Search AI (MCTS)
-
-Modern MCTS-based AI with multithreading support:
-
- **Advantages**: Multithreaded, better performance on modern CPUs, anytime algorithm
- **Use cases**: Performance testing, scenarios requiring fast decisions
- **Performance**: Multithreaded, adaptive depth based on time budget
-
-### Switching Between Algorithms
-
-The algorithm is selected at **runtime** via the ShardokAIClient constructor:
-
-```cpp
-// Using Iterative Deepening AI (default)
-ShardokAIClient client(playerId, isDefender, hexMap, settings);
-// OR explicitly:
-ShardokAIClient client(playerId, isDefender, hexMap, settings, AIAlgorithmType::ITERATIVE_DEEPENING);
-
-// Using MCTS AI
-ShardokAIClient client(playerId, isDefender, hexMap, settings, AIAlgorithmType::MCTS);
-```
-
-```bash
-# Build the server (includes both AI algorithms)
-bazel build //src/main/cpp/net/eagle0/shardok:shardok-server
-
-# Test both algorithms
-bazel test //src/test/cpp/net/eagle0/shardok/ai:ai_iterative_deepening_test
-bazel test //src/test/cpp/net/eagle0/shardok/ai:ai_mcts_test  # If available
-
-# Performance tests
-./scripts/ai_perf_test.sh  # Uses whatever algorithm the server is configured to use
-```
-
-Both implementations are compatible with all existing interfaces and produce the same `SearchResult` structure.
-
-**Note**: Both implementations are documented in `src/main/cpp/net/eagle0/shardok/ai/AI_SCORING_SYSTEM.md`, including
-recommendations for improving MCTS randomness handling.
-
-The AI algorithm selection is made at runtime when creating ShardokAIClient instances, allowing different AI strategies
-to be used for different players or game situations within the same server process.
-
-## Language-Specific Patterns
-
-**Scala (Strategic Layer):**
-
- Use `EngineImpl.scala` for core game logic modifications
- Follow event sourcing pattern - all changes through immutable actions
- gRPC streaming for real-time client updates via `EagleServiceImpl.scala`
- LLM integration in `/common/llm_integration/` for narrative generation
-
-**C++ (Tactical Layer):**
-
- Performance-critical combat in `ShardokEngine.hpp/.cpp`
- FlatBuffers for efficient serialization in `/flatbuffer/` directory
- AI systems in `/ai/` subdirectory with pluggable strategy selectors
- Extensive unit testing with Google Test framework
-
-**Protocol Buffers:**
-
- Three-layer structure: `api/` (client), `internal/` (server), `views/` (projections)
- Use `shardok_internal_interface.proto` for Eagle-Shardok communication
- Maintain backward compatibility when modifying existing messages
-
-**C# (Unity Client):**
-
- Located in `/src/main/csharp/net/eagle0/clients/unity/eagle0/`
- Uses Unity 6 (6000.4.10f1) with comprehensive protobuf integration (100+ .proto files)
- Key components: `EagleConnection.cs` (gRPC client), `EagleGameController.cs` (main game logic)
- Real-time bidirectional streaming with server via `PersistentClientConnection.cs`
- Strategic map UI in `Assets/Eagle/`, tactical battle UI in `Assets/Shardok/`
- Seamless transition between strategic gameplay and hex-based tactical combat
- **NEVER add defensive null checks on Unity Inspector fields** - these hide configuration bugs. If a field isn't
-  linked in the editor, it should throw a NullReferenceException so the problem is immediately obvious. Silently
-  skipping code when a required field is null makes bugs harder to find.
-
-**Go (Build Tools):**
-
- Build automation and code generation utilities
- AWS S3 integration for deployment artifacts
-
-## Testing Strategy
-
- Comprehensive unit tests for both Scala and C++ components
- Integration tests for Eagle-Shardok communication
- Map validation tests ensure game content integrity
- Use `GameSettings_test_utils.cpp` and `ShardokEngineBasedTestData.cpp` for C++ test helpers
-
-### Scala Testing Patterns
-
-**Use `inside()` instead of `asInstanceOf` for type matching in tests:**
-
-Never use `asInstanceOf` in tests. Instead, use ScalaTest's `inside()` pattern for safe type matching:
-
-```scala
-// BAD - don't do this
-val changedHero = result.changedHeroes.head.asInstanceOf[ChangedHeroC]
-changedHero.heroId shouldBe 19
-
-// GOOD - use inside() pattern
-import org.scalatest.Inside.inside
-
-inside(result.changedHeroes.head) { case changedHero: ChangedHeroC =>
-  changedHero.heroId shouldBe 19
-  changedHero.vigorChange shouldBe StatDelta(17.2)
-}
-```
-
-The `inside()` pattern:
- Provides better error messages when the type doesn't match
- Is idiomatic ScalaTest
- Works with pattern matching for more complex assertions
-
-## Performance Testing
-
-When making performance-related changes to the AI or engine:
-
-```bash
-# 1. Commit your changes to a feature branch
-git checkout -b performance-improvement-feature
-git add . && git commit -m "Implement performance improvement"
-
-# 2. Run performance tests multiple times on your branch to reduce noise
-for i in 1 2 3; do 
-    echo "=== Run $i ===" 
-    ./scripts/ai_perf_test.sh 2>&1 | grep -A 20 "AI Search Performance Summary" 
-done
-# Save or note the results
-
-# 3. Switch to main branch and run the same tests
-git checkout main
-for i in 1 2 3; do 
-    echo "=== Run $i ===" 
-    ./scripts/ai_perf_test.sh 2>&1 | grep -A 20 "AI Search Performance Summary" 
-done
-
-# 4. Compare the results between your branch and main
-# Key metrics to compare:
-# - Commands evaluated at each depth (e.g., "Depth 3: 169/523 commands")
-# - Average search depth achieved
-# - Completion rates at each depth
-```
-
-**Important notes:**
-
- Run tests multiple times (3-5) to account for performance variance
- Focus on commands evaluated at each depth rather than total commands
- Commands at different depths aren't directly comparable (depth 3 is more valuable than depth 2)
- **Always test performance changes** - what seems like an optimization may sometimes have unexpected overhead or
-  behavior changes.
-
-## Troubleshooting Scala Build Errors
-
-### MissingType Errors
-
-When you see errors like:
-```
-dotty.tools.dotc.core.MissingType: Cannot resolve reference to type net.eagle0.eagle.internal.game_state.type.GameState
-```
-
-**This is NOT a Scala compiler crash.** This is a missing dependency in BUILD.bazel.
-
-**How to fix:**
-1. Identify the missing type from the error message (e.g., `game_state.GameState`)
-2. Find the Bazel target that provides this type (e.g., `//src/main/protobuf/net/eagle0/eagle/internal:game_state_scala_proto`)
-3. Add it to the `deps` of the failing target
-4. If the type appears in a public method signature, also add it to `exports` so downstream targets can see it
-
-**Common pattern:** When adding a method to a class that takes or returns a proto type, the proto dependency often needs to be added to both `deps` AND `exports`.
-
-### Bazel Clean
-
-**NEVER run `bazel clean` without asking first.** It rarely fixes actual issues and wastes significant rebuild time. The issues that seem like they need `bazel clean` are usually:
- Missing imports in Scala code
- Missing dependencies in BUILD.bazel
- Missing exports for types used in public signatures
-
-## Game Content
-
-**Maps:** `.e0mj` files in `/src/main/resources/net/eagle0/shardok/maps/`
-**Configuration:** Game parameters in `/src/main/resources/net/eagle0/eagle/game_parameters.json`
-**Data Files:** TSV format for battalions, heroes, and other game data
-
-**NEVER modify hero names.** Do not change names in `heroes.tsv`, `generated_heroes`, or any other hero data files. The names are carefully chosen and are not to be altered.
-
-## Deployment
-
- Bazel handles multi-language builds and dependencies
- CI/CD via GitHub Actions with platform-specific build scripts in `/ci/github_actions/`
- Docker containerization available via `ci/eagle_run.Dockerfile`
- Always run "bazel run //:gazelle" after editing any BUILD.bazel files
- *ALWAYS ALWAYS* run "bazel run gazelle" after any change that modifies a BUILD.bazel file
@@ -4,11 +4,9 @@

 **NEVER chain bash commands.** Do not use `&&`, `||`, or `;` to combine commands. Each command must be a separate Bash tool call. Use parallel tool calls when commands are independent.

-**NEVER prefix a command with `cd`.** Run `git`, `gh`, `bazel`, etc. directly from whatever the cwd already is. The cwd is a subdirectory of the worktree; `git`/`gh` find the repo via `.git` discovery and `bazel` finds the workspace via MODULE.bazel discovery — walking up from a subdir works fine. The shell resets cwd after every command anyway, so a `cd` never persists. (And `cd <dir> && <cmd>` violates the no-chaining rule above and forces an approval prompt every time.) Only `cd` in the rare case a tool genuinely cannot locate its root, and even then as its own Bash call, never a chain.
-
 ## CRITICAL GIT RULES (NEVER VIOLATE)

-**NEVER use `git -C`.** Just run `git` directly from the cwd — it finds the repo via `.git` discovery. Do not `cd` to the repo root either (see the no-`cd` bash rule above).
+**NEVER use `git -C`.** Instead, cd to the repo root and run git commands from there.

 **NEVER push directly to main/master.** No exceptions. Not for "small changes." Not for docs. Not ever.

@@ -228,7 +226,7 @@ to be used for different players or game situations within the same server proce
 **C# (Unity Client):**

 - Located in `/src/main/csharp/net/eagle0/clients/unity/eagle0/`
- Uses Unity 6 (6000.4.10f1) with comprehensive protobuf integration (100+ .proto files)
+- Uses Unity 6 (6000.0.32f1) with comprehensive protobuf integration (100+ .proto files)
 - Key components: `EagleConnection.cs` (gRPC client), `EagleGameController.cs` (main game logic)
 - Real-time bidirectional streaming with server via `PersistentClientConnection.cs`
 - Strategic map UI in `Assets/Eagle/`, tactical battle UI in `Assets/Shardok/`
@@ -352,4 +350,4 @@ dotty.tools.dotc.core.MissingType: Cannot resolve reference to type net.eagle0.e
 - CI/CD via GitHub Actions with platform-specific build scripts in `/ci/github_actions/`
 - Docker containerization available via `ci/eagle_run.Dockerfile`
 - Always run "bazel run //:gazelle" after editing any BUILD.bazel files
- *ALWAYS ALWAYS* run "bazel run gazelle" after any change that modifies a BUILD.bazel file
+- *ALWAYS ALWAYS* run "bazel run gazelle" after any change that modifies a BUILD.bazel file
@@ -3,29 +3,25 @@ module(name = "net_eagle0")
 # Version constants
 SCALA_VERSION = "3.7.4"

-NETTY_VERSION = "4.1.135.Final"
+NETTY_VERSION = "4.1.127.Final"

-SCALAPB_VERSION = "1.0.0-alpha.5"
+SCALAPB_VERSION = "1.0.0-alpha.1"

-SCALAPB_JSON4S_VERSION = "1.0.0-alpha.1"
-
-AWS_SDK_VERSION = "2.46.7"
-
-SLF4J_VERSION = "2.0.18"
+AWS_SDK_VERSION = "2.41.18"

 #
 # Core Build Tools
 #

 bazel_dep(name = "bazel_skylib", version = "1.9.0")
-bazel_dep(name = "platforms", version = "1.1.0")
+bazel_dep(name = "platforms", version = "1.0.0")
 bazel_dep(name = "rules_pkg", version = "1.2.0")

 #
 # Language Support - Scala
 #

-bazel_dep(name = "rules_scala", version = "7.2.5")
+bazel_dep(name = "rules_scala", version = "7.2.1")

 scala_config = use_extension(
    "@rules_scala//scala/extensions:config.bzl",
@@ -45,21 +41,21 @@ scala_deps.scala_proto()
 # Language Support - C++
 #

-bazel_dep(name = "toolchains_llvm", version = "1.7.0")
+bazel_dep(name = "toolchains_llvm", version = "1.6.0")

 llvm = use_extension("@toolchains_llvm//toolchain/extensions:llvm.bzl", "llvm")

 # Native toolchain (macOS -> macOS, Linux -> Linux)
 llvm.toolchain(
    name = "llvm_toolchain",
-    llvm_version = "20.1.8",
+    llvm_version = "20.1.4",
 )

 # Cross-compilation toolchain (macOS -> Linux x86_64)
 # Uses the same LLVM distribution but with a Linux sysroot
 llvm.toolchain(
    name = "llvm_toolchain_linux",
-    llvm_version = "20.1.8",
+    llvm_version = "20.1.4",
 )

 # Linux x86_64 sysroot for cross-compilation
@@ -72,7 +68,7 @@ llvm.sysroot(
 # Cross-compilation toolchain (macOS -> Linux ARM64)
 llvm.toolchain(
    name = "llvm_toolchain_linux_arm64",
-    llvm_version = "20.1.8",
+    llvm_version = "20.1.4",
 )

 # Linux ARM64 sysroot for cross-compilation
@@ -81,6 +77,7 @@ llvm.sysroot(
    label = "@linux_sysroot_arm64//sysroot",
    targets = ["linux-aarch64"],
 )
+
 use_repo(llvm, "llvm_toolchain", "llvm_toolchain_linux", "llvm_toolchain_linux_arm64")

 # Download the Linux sysroots (Ubuntu 24.04 Noble for C++23 support)
@@ -106,11 +103,11 @@ sysroot(
 # Language Support - Go
 #

-bazel_dep(name = "rules_go", version = "0.61.1", repo_name = "io_bazel_rules_go")
-bazel_dep(name = "gazelle", version = "0.51.3", repo_name = "bazel_gazelle")
+bazel_dep(name = "rules_go", version = "0.59.0", repo_name = "io_bazel_rules_go")
+bazel_dep(name = "gazelle", version = "0.47.0", repo_name = "bazel_gazelle")

 go_sdk = use_extension("@io_bazel_rules_go//go:extensions.bzl", "go_sdk")
-go_sdk.download(version = "1.25.11")
+go_sdk.download(version = "1.23.3")
 use_repo(go_sdk, "go_default_sdk")

 go_deps = use_extension("@bazel_gazelle//:extensions.bzl", "go_deps")
@@ -133,7 +130,7 @@ use_repo(
 # Language Support - Rust
 #

-bazel_dep(name = "rules_rust", version = "0.70.0")
+bazel_dep(name = "rules_rust", version = "0.68.1")

 rust = use_extension("@rules_rust//rust:extensions.bzl", "rust")
 rust.toolchain(edition = "2021")
@@ -153,7 +150,7 @@ use_repo(crate, "map_generator_crates")
 # Platform Support - Apple/iOS
 #

-bazel_dep(name = "apple_support", version = "1.24.5", repo_name = "build_bazel_apple_support")
+bazel_dep(name = "apple_support", version = "1.24.1", repo_name = "build_bazel_apple_support")
 bazel_dep(name = "rules_apple", version = "4.3.3", repo_name = "build_bazel_rules_apple")

 # Register Apple CC toolchain for Objective-C compilation
@@ -167,7 +164,7 @@ use_repo(apple_cc_configure, "local_config_apple_cc", "local_config_apple_cc_too
 # Protocol Buffers & RPC
 #

-bazel_dep(name = "protobuf", version = "33.6", repo_name = "com_google_protobuf")
+bazel_dep(name = "protobuf", version = "33.5", repo_name = "com_google_protobuf")

 # Use pre-built protoc binaries instead of compiling from source.
 # See: https://protobuf.dev/reference/cpp/cpp-generated/#invocation
@@ -186,7 +183,7 @@ use_repo(
 )

 bazel_dep(name = "rules_proto", version = "7.1.0")
-bazel_dep(name = "rules_cc", version = "0.2.19")
+bazel_dep(name = "rules_cc", version = "0.2.14")
 bazel_dep(name = "grpc", version = "1.78.0")

 # Patch grpc 1.78.0 to fix rules_python incompatibility: grpc requests Python
@@ -200,7 +197,7 @@ single_version_override(
    version = "1.78.0",
 )

-bazel_dep(name = "grpc-java", version = "1.78.0.bcr.1")
+bazel_dep(name = "grpc-java", version = "1.78.0")
 bazel_dep(name = "rules_proto_grpc_csharp", version = "5.8.0")
 bazel_dep(name = "flatbuffers", version = "25.12.19")

@@ -208,13 +205,13 @@ bazel_dep(name = "flatbuffers", version = "25.12.19")
 # Testing
 #

-bazel_dep(name = "googletest", version = "1.17.0.bcr.2")
+bazel_dep(name = "googletest", version = "1.17.0")

 #
 # Container Images (OCI)
 #

-bazel_dep(name = "rules_oci", version = "2.3.0")
+bazel_dep(name = "rules_oci", version = "2.2.7")
 bazel_dep(name = "aspect_bazel_lib", version = "2.22.5")

 oci = use_extension("@rules_oci//oci:extensions.bzl", "oci")
@@ -264,7 +261,7 @@ use_repo(oci, "alpine_linux", "alpine_linux_linux_amd64", "eclipse_temurin_25_jd
 # Java/Scala Dependencies
 #

-bazel_dep(name = "rules_java", version = "9.6.1")
+bazel_dep(name = "rules_java", version = "9.3.0")
 bazel_dep(name = "rules_jvm_external", version = "6.10")

 maven = use_extension("@rules_jvm_external//:extensions.bzl", "maven")
@@ -284,7 +281,7 @@ maven.install(

        # ScalaPB
        "com.thesamet.scalapb:lenses_3:%s" % SCALAPB_VERSION,
-        "com.thesamet.scalapb:scalapb-json4s_3:%s" % SCALAPB_JSON4S_VERSION,
+        "com.thesamet.scalapb:scalapb-json4s_3:%s" % SCALAPB_VERSION,
        "com.thesamet.scalapb:scalapb-runtime_3:%s" % SCALAPB_VERSION,
        "com.thesamet.scalapb:scalapb-runtime-grpc_3:%s" % SCALAPB_VERSION,
        "com.thesamet.scalapb:compilerplugin_3:%s" % SCALAPB_VERSION,
@@ -296,10 +293,7 @@ maven.install(
        "org.json4s:json4s-native_3:4.1.0-M8",

        # Testing
-        "org.scalamock:scalamock_3:7.5.5",
-
-        # Developer tools
-        "org.scalameta:scalafmt-cli_2.13:3.11.1",
+        "org.scalamock:scalamock_3:7.4.1",

        # AWS SDK
        "software.amazon.awssdk:s3-transfer-manager:%s" % AWS_SDK_VERSION,
@@ -311,69 +305,58 @@ maven.install(
        "software.amazon.awssdk:http-client-spi:%s" % AWS_SDK_VERSION,

        # AWS Lambda
-        "com.amazonaws:aws-lambda-java-core:1.4.0",
-        "com.amazonaws:aws-lambda-java-events:3.16.1",
+        "com.amazonaws:aws-lambda-java-core:1.2.3",
+        "com.amazonaws:aws-lambda-java-events:3.13.0",

        # Logging
-        "org.slf4j:slf4j-api:%s" % SLF4J_VERSION,
-        "org.slf4j:slf4j-simple:%s" % SLF4J_VERSION,
+        "org.slf4j:slf4j-api:2.0.16",
+        "org.slf4j:slf4j-simple:2.0.16",

        # Other
        "org.reactivestreams:reactive-streams:1.0.4",
        "javax.xml.bind:jaxb-api:2.3.1",

        # OkHttp (for SSE with read timeout support, OAuth HTTP calls)
-        "com.squareup.okhttp3:okhttp-jvm:5.4.0",
-        "com.squareup.okhttp3:okhttp-sse:5.4.0",
+        "com.squareup.okhttp3:okhttp:4.12.0",
+        "com.squareup.okhttp3:okhttp-sse:4.12.0",

        # JWT (for OAuth token handling)
-        "com.nimbusds:nimbus-jose-jwt:10.9.1",
+        "com.nimbusds:nimbus-jose-jwt:9.37.3",

        # Error tracking
-        "io.sentry:sentry:8.43.2",
+        "io.sentry:sentry:8.31.0",

        # SQLite for client text storage
-        "org.xerial:sqlite-jdbc:3.53.2.0",
-
-        # Postgres for production history benchmarking and migration work
-        "org.postgresql:postgresql:42.7.11",
+        "org.xerial:sqlite-jdbc:3.46.1.0",
    ],
    duplicate_version_warning = "error",
    fail_if_repin_required = True,
    lock_file = "//:maven_install.json",
    repositories = [
-        "https://repo.maven.apache.org/maven2",
+        "https://repo1.maven.org/maven2",
    ],
 )

-# json4s' Scala 3 reflection path can load scala.quoted.staging at runtime, but
-# its Maven metadata does not bring the staging jar transitively.
-maven.artifact(
-    artifact = "scala3-staging_3",
-    exclusions = ["org.scala-lang:scala3-compiler_3"],
-    group = "org.scala-lang",
-    version = SCALA_VERSION,
-)
-
 # Force specific versions for dependencies with conflicts between grpc-java and protobuf
 maven.artifact(
    artifact = "gson",
    force_version = True,
    group = "com.google.code.gson",
-    version = "2.14.0",
+    version = "2.11.0",
 )
 maven.artifact(
    artifact = "error_prone_annotations",
    force_version = True,
    group = "com.google.errorprone",
-    version = "2.50.0",
+    version = "2.30.0",
 )
 maven.artifact(
    artifact = "guava",
    force_version = True,
    group = "com.google.guava",
-    version = "33.6.0-android",
+    version = "33.3.1-android",
 )
+
 use_repo(maven, "maven", "unpinned_maven")

 #
@@ -381,7 +364,6 @@ use_repo(maven, "maven", "unpinned_maven")
 #

 http_archive = use_repo_rule("@bazel_tools//tools/build_defs/repo:http.bzl", "http_archive")
-
 http_file = use_repo_rule("@bazel_tools//tools/build_defs/repo:http.bzl", "http_file")

 # GTL (for parallel_hashmap)
@@ -434,24 +416,24 @@ http_archive(
 # Fallback: busybox.net (can be unreliable/slow)
 http_file(
    name = "busybox_x86_64",
-    downloaded_file_path = "busybox",
-    executable = True,
    sha256 = "6e123e7f3202a8c1e9b1f94d8941580a25135382b99e8d3e34fb858bba311348",
    urls = [
        "https://eagle0-sysroot.sfo3.digitaloceanspaces.com/busybox/busybox-1.35.0-x86_64-linux-musl",
        "https://busybox.net/downloads/binaries/1.35.0-x86_64-linux-musl/busybox",
    ],
+    downloaded_file_path = "busybox",
+    executable = True,
 )

 http_file(
    name = "busybox_aarch64",
-    downloaded_file_path = "busybox",
-    executable = True,
    sha256 = "141adb1b625a6f44c4b114f76b4387b4ea4f7ab802b88eb40e0d2f6adcccb1c3",
    urls = [
        # TODO: Upload aarch64 binary to GitHub release when needed
        "https://busybox.net/downloads/binaries/1.35.0-aarch64-linux-musl/busybox",
    ],
+    downloaded_file_path = "busybox",
+    executable = True,
 )

 # LLVM MinGW toolchain for Windows cross-compilation from macOS
@@ -1,6 +1,5 @@
 load("@rules_oci//oci:defs.bzl", "oci_image", "oci_load", "oci_push")
 load("@rules_pkg//pkg:tar.bzl", "pkg_tar")
-load("//ci:jar_split.bzl", "jar_split")

 #
 # Deployment artifacts (tools needed on the host, not in containers)
@@ -51,37 +50,11 @@ pkg_tar(
 # Push:  bazel run //ci:eagle_server_push
 #

-# Split the Eagle server runtime classpath into a stable third-party layer
-# (Scala stdlib, gRPC, Netty, ScalaPB, AWS SDK, ...) and a small first-party
-# layer that changes every commit. Most pushes then only re-upload the small
-# app layer instead of the whole ~150-300MB fat JAR.
-jar_split(
-    name = "eagle_server_jars",
-    binary = "//src/main/scala/net/eagle0/eagle:eagle_server",
-)
-
-filegroup(
-    name = "eagle_server_deps_jars",
-    srcs = [":eagle_server_jars"],
-    output_group = "deps",
-)
-
-filegroup(
-    name = "eagle_server_app_jars",
-    srcs = [":eagle_server_jars"],
-    output_group = "app",
-)
-
+# Package the deploy JAR
 pkg_tar(
-    name = "eagle_server_deps_layer",
-    srcs = [":eagle_server_deps_jars"],
-    package_dir = "/app/lib/deps",
-)
-
-pkg_tar(
-    name = "eagle_server_app_layer",
-    srcs = [":eagle_server_app_jars"],
-    package_dir = "/app/lib/app",
+    name = "eagle_server_jar_layer",
+    srcs = ["//src/main/scala/net/eagle0/eagle:eagle_server_deploy.jar"],
+    package_dir = "/app",
 )

 # Package the game resources needed at runtime
@@ -109,11 +82,8 @@ oci_image(
        "-XX:+UnlockDiagnosticVMOptions",
        "-XX:+DebugNonSafepoints",  # Required for JFR to see through inlined methods
        "-XX:FlightRecorderOptions=stackdepth=256",
-        # Classpath glob is expanded by the JVM itself (exec-form, no shell).
-        # Deps dir first keeps third-party precedence; app last shadows nothing.
-        "-cp",
-        "/app/lib/deps/*:/app/lib/app/*",
-        "net.eagle0.eagle.Main",
+        "-jar",
+        "/app/eagle_server_deploy.jar",
    ],
    env = {
        "JAVA_OPTS": "-Xmx2g -XX:+UseG1GC",
@@ -121,8 +91,7 @@ oci_image(
    exposed_ports = ["40032/tcp"],
    tars = [
        ":busybox_layer",
-        ":eagle_server_deps_layer",
-        ":eagle_server_app_layer",
+        ":eagle_server_jar_layer",
        ":eagle_resources_layer",
    ],
    workdir = "/app",
@@ -39,12 +39,6 @@ if [ $UNITY_EXIT_CODE -ne 0 ]; then
    exit $UNITY_EXIT_CODE
 fi

-if [ ! -f "$LOG_PATH" ]; then
-    echo ""
-    echo "ERROR: Unity exited successfully but did not write an editor log at $LOG_PATH"
-    exit 1
-fi
-
 # Fail the build if any prefab references are broken — this produces a player
 # that launches but has null Inspector fields, which is hard to debug.
 if grep -q "Missing Prefab" "$LOG_PATH"; then
@@ -39,12 +39,6 @@ if [ $UNITY_EXIT_CODE -ne 0 ]; then
    exit $UNITY_EXIT_CODE
 fi

-if [ ! -f "$LOG_PATH" ]; then
-    echo ""
-    echo "ERROR: Unity exited successfully but did not write an editor log at $LOG_PATH"
-    exit 1
-fi
-
 # Fail the build if any prefab references are broken — this produces a player
 # that launches but has null Inspector fields, which is hard to debug.
 if grep -q "Missing Prefab" "$LOG_PATH"; then
@@ -1,82 +0,0 @@
-#!/usr/bin/env bash
-
-set -euo pipefail
-
-COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-export PATH="${COMMON_PATHS}:${PATH}"
-
-write_developer_dir_override() {
-    if [ "$(uname -s)" != "Darwin" ]; then
-        return
-    fi
-
-    local developer_dir=""
-    if [ -d "/Applications/Xcode.app/Contents/Developer" ]; then
-        developer_dir="/Applications/Xcode.app/Contents/Developer"
-    elif [ -d "/Library/Developer/CommandLineTools" ]; then
-        developer_dir="/Library/Developer/CommandLineTools"
-    else
-        developer_dir="$(xcode-select -p 2>/dev/null || true)"
-    fi
-
-    if [ -z "$developer_dir" ] || [ ! -d "$developer_dir" ]; then
-        echo "ERROR: no usable Apple developer directory found"
-        echo "Install Command Line Tools with: xcode-select --install"
-        echo "Full Xcode is only required for workflows that build/sign Apple apps."
-        exit 1
-    fi
-
-    export DEVELOPER_DIR="$developer_dir"
-    if [ -f .bazelrc.local ] && [ -z "${CI:-}" ]; then
-        echo "Keeping existing .bazelrc.local outside CI"
-    else
-        {
-            echo "# Generated by ci/github_actions/ensure_bazel_installed.sh; do not edit."
-            printf 'common:macos --repo_env=DEVELOPER_DIR=%s\n' "$developer_dir"
-        } > .bazelrc.local
-    fi
-
-    if [ -n "${GITHUB_ENV:-}" ]; then
-        echo "DEVELOPER_DIR=$developer_dir" >> "$GITHUB_ENV"
-    fi
-
-    echo "Using Apple developer directory at $developer_dir"
-}
-
-write_developer_dir_override
-
-if [ -n "${GITHUB_PATH:-}" ]; then
-    for path in /opt/homebrew/bin /usr/local/bin; do
-        if [ -d "$path" ]; then
-            echo "$path" >> "$GITHUB_PATH"
-        fi
-    done
-fi
-
-if command -v bazel >/dev/null 2>&1; then
-    echo "Using bazel at $(command -v bazel)"
-    bazel --version
-    exit 0
-fi
-
-if ! command -v bazelisk >/dev/null 2>&1; then
-    if ! command -v brew >/dev/null 2>&1; then
-        echo "ERROR: neither bazel nor bazelisk is on PATH, and Homebrew is unavailable"
-        exit 1
-    fi
-
-    echo "Installing bazelisk with Homebrew"
-    brew install bazelisk
-fi
-
-BAZEL_BIN_DIR="${RUNNER_TEMP:-/tmp}/bazel-bin"
-mkdir -p "$BAZEL_BIN_DIR"
-ln -sf "$(command -v bazelisk)" "$BAZEL_BIN_DIR/bazel"
-
-if [ -n "${GITHUB_PATH:-}" ]; then
-    echo "$BAZEL_BIN_DIR" >> "$GITHUB_PATH"
-fi
-
-export PATH="$BAZEL_BIN_DIR:$PATH"
-echo "Using bazelisk as bazel at $BAZEL_BIN_DIR/bazel"
-bazel --version
@@ -34,7 +34,6 @@ echo "Platform: ${PLATFORM}"
 # Check if Unity is installed and has required modules
 check_modules_installed() {
    local unity_path="${UNITY_INSTALL_PATH}/${UNITY_VERSION}"
-    local unity_app_contents_path="${unity_path}/Unity.app/Contents"

    if [ ! -d "$unity_path" ]; then
        return 1
@@ -50,8 +49,7 @@ check_modules_installed() {
            ;;
        mac)
            # Mac IL2CPP module
-            if [ ! -d "${unity_app_contents_path}/PlaybackEngines/MacStandaloneSupport/Source/Player/MacPlayer/GameAssembly.xcodeproj" ] && \
-               [ ! -d "${unity_path}/PlaybackEngines/MacStandaloneSupport/Variations/macos_development_il2cpp" ] && \
+            if [ ! -d "${unity_path}/PlaybackEngines/MacStandaloneSupport/Variations/macos_development_il2cpp" ] && \
               [ ! -d "${unity_path}/PlaybackEngines/MacStandaloneSupport" ]; then
                echo "✗ Mac IL2CPP module not installed for Unity ${UNITY_VERSION}"
                return 1
@@ -71,8 +69,7 @@ check_modules_installed() {
                echo "✗ iOS module missing"
                missing=1
            fi
-            if [ ! -d "${unity_app_contents_path}/PlaybackEngines/MacStandaloneSupport/Source/Player/MacPlayer/GameAssembly.xcodeproj" ] && \
-               [ ! -d "${unity_path}/PlaybackEngines/MacStandaloneSupport" ]; then
+            if [ ! -d "${unity_path}/PlaybackEngines/MacStandaloneSupport" ]; then
                echo "✗ Mac module missing"
                missing=1
            fi
@@ -194,10 +191,8 @@ echo "Running: ${INSTALL_CMD[*]}"
 echo ""

 # Capture output to check for "already installed" messages
-set +e
-OUTPUT=$("${INSTALL_CMD[@]}" 2>&1)
+OUTPUT=$("${INSTALL_CMD[@]}" 2>&1) || true
 EXIT_CODE=$?
-set -e
 echo "$OUTPUT"

 # Check if modules are already installed (Unity Hub returns error but modules are present)
@@ -224,13 +219,13 @@ else
    exit 1
 fi

-# Verify installation and requested modules
+# Verify installation
 echo ""
-if check_modules_installed; then
-    echo "✓ Verified: Unity ${UNITY_VERSION} is now installed with ${PLATFORM} support"
+if [ -d "${UNITY_INSTALL_PATH}/${UNITY_VERSION}" ]; then
+    echo "✓ Verified: Unity ${UNITY_VERSION} is now installed"
    exit 0
 else
-    echo "✗ Unity ${UNITY_VERSION} installation with ${PLATFORM} support could not be verified"
+    echo "✗ Unity ${UNITY_VERSION} installation could not be verified"
    echo "  Expected path: ${UNITY_INSTALL_PATH}/${UNITY_VERSION}"
    echo ""
    echo "The installation may still be in progress, or may require manual intervention."
@@ -11,46 +11,16 @@

 set -euo pipefail

-COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-export PATH="${COMMON_PATHS}:${PATH}"
-
 MAX_ATTEMPTS=5
 RETRY_DELAY=15

-if [ -n "${GITHUB_PATH:-}" ]; then
-    for path in /opt/homebrew/bin /usr/local/bin; do
-        if [ -d "$path" ]; then
-            echo "$path" >> "$GITHUB_PATH"
-        fi
-    done
-fi
-
-if ! git lfs version >/dev/null 2>&1; then
-    if ! command -v brew >/dev/null 2>&1; then
-        echo "ERROR: git-lfs is unavailable, and Homebrew is not on PATH"
-        exit 1
-    fi
-
-    echo "Installing git-lfs with Homebrew"
-    brew install git-lfs
-fi
-
-echo "Using git-lfs at $(command -v git-lfs)"
-git lfs version
-
-git lfs install --force
+git lfs install

 echo "LFS objects before pull:"
 git lfs ls-files | wc -l

-GIT_LFS_PULL=(git lfs pull)
-if [ -n "${GITHUB_TOKEN:-}" ]; then
-    BASIC_AUTH=$(printf 'x-access-token:%s' "$GITHUB_TOKEN" | base64 | tr -d '\n')
-    GIT_LFS_PULL=(git -c "http.https://github.com/.extraheader=AUTHORIZATION: basic $BASIC_AUTH" lfs pull)
-fi
-
 for i in $(seq 1 $MAX_ATTEMPTS); do
-    if "${GIT_LFS_PULL[@]}" "$@"; then
+    if git lfs pull "$@"; then
        echo "LFS objects after pull:"
        git lfs ls-files | wc -l
        exit 0
@@ -1,56 +0,0 @@
-#!/usr/bin/env bash
-
-set -euxo pipefail
-
-# Read Unity version from project file
-UNITY_VERSION=$(grep "m_EditorVersion:" src/main/csharp/net/eagle0/clients/unity/eagle0/ProjectSettings/ProjectVersion.txt | head -1 | sed 's/m_EditorVersion: //')
-
-# Use runner-specific build directory if EAGLE0_BUILD_DIR is set, otherwise default
-BUILD_BASE="${EAGLE0_BUILD_DIR:-/tmp/eagle0}"
-
-WORKSPACE=$(pwd)
-UNITY_INSTALL_PATH="/Applications/Unity/Hub/Editor"
-LOG_PATH="${BUILD_BASE}/editor_editmode_tests.log"
-RESULTS_PATH="${BUILD_BASE}/editmode-test-results.xml"
-
-mkdir -p "$BUILD_BASE"
-
-echo "Building protos"
-./scripts/build_protos.sh
-
-echo "Running Unity EditMode tests"
-
-set +e
-"${UNITY_INSTALL_PATH}/${UNITY_VERSION}/Unity.app/Contents/MacOS/Unity" \
-   -nographics \
-   -batchmode \
-   -quit \
-   -executeMethod eagle0.Tests.DynamicTextBindingTestRunner.Run \
-   -testResults "$RESULTS_PATH" \
-   -logFile "$LOG_PATH" \
-   -projectPath "$WORKSPACE/src/main/csharp/net/eagle0/clients/unity/eagle0"
-UNITY_EXIT_CODE=$?
-set -e
-
-if [ $UNITY_EXIT_CODE -ne 0 ]; then
-    echo ""
-    echo "Unity EditMode tests failed with exit code $UNITY_EXIT_CODE"
-    echo "=== Unity EditMode Test Log (last 200 lines) ==="
-    tail -200 "$LOG_PATH" || echo "Could not read log file at $LOG_PATH"
-    echo "=== End of Unity EditMode Test Log ==="
-    exit $UNITY_EXIT_CODE
-fi
-
-if [ ! -f "$LOG_PATH" ]; then
-    echo ""
-    echo "ERROR: Unity exited successfully but did not write an editor log at $LOG_PATH"
-    exit 1
-fi
-
-if [ ! -f "$RESULTS_PATH" ]; then
-    echo ""
-    echo "ERROR: Unity exited successfully but did not write test results at $RESULTS_PATH"
-    exit 1
-fi
-
-echo "Unity EditMode tests complete"
@@ -19,18 +19,6 @@ SERVER_DATA="$UNITY_PROJECT/ServerData/$BUILD_TARGET"
 DO_ENDPOINT="https://sfo3.digitaloceanspaces.com"
 DO_BUCKET="eagle0-assets"

-COMMON_PATHS="/opt/homebrew/bin:/usr/local/bin:/usr/bin:/bin:/usr/sbin:/sbin"
-export PATH="${COMMON_PATHS}:${PATH}"
-
-if ! command -v aws >/dev/null 2>&1; then
-    if command -v brew >/dev/null 2>&1; then
-        brew install awscli
-    else
-        echo "ERROR: aws CLI is unavailable, and Homebrew is not on PATH"
-        exit 1
-    fi
-fi
-
 if [ ! -d "$SERVER_DATA" ]; then
    echo "No Addressables bundles found at $SERVER_DATA"
    echo "Skipping upload (this is expected if Addressables are bundled locally)"
@@ -1,57 +0,0 @@
-"""Split a JVM binary's runtime classpath into first-party and third-party jar sets.
-
-This exists so the Docker image can place rarely-changing third-party jars in a
-lower OCI layer and frequently-changing first-party jars in a small top layer,
-so most pushes only re-upload the small layer.
-"""
-
-def _unique_name(jar):
-    # short_path is unique per jar and stable across commits (it depends only on
-    # the jar's own package/coordinate, not on unrelated targets), so the deps
-    # layer's tar entries stay byte-identical and its blob digest stays cached.
-    path = jar.short_path
-    if path.startswith("../"):
-        path = path[3:]
-    return path.replace("+", "_").replace("/", "_").replace("~", "_")
-
-def _impl(ctx):
-    info = ctx.attr.binary[JavaInfo]
-
-    app = []
-    deps = []
-    seen = {}
-    for jar in sorted(info.transitive_runtime_jars.to_list(), key = lambda f: f.path):
-        workspace = jar.owner.workspace_name if jar.owner else ""
-        bucket = "app" if workspace == "" else "deps"
-        out_name = _unique_name(jar)
-
-        key = bucket + "/" + out_name
-        if key in seen:
-            fail("jar_split: duplicate output name %r for %s and %s" % (
-                key,
-                seen[key],
-                jar.path,
-            ))
-        seen[key] = jar.path
-
-        link = ctx.actions.declare_file(ctx.label.name + "/" + key)
-        ctx.actions.symlink(output = link, target_file = jar)
-        (app if bucket == "app" else deps).append(link)
-
-    return [
-        DefaultInfo(files = depset(app + deps)),
-        OutputGroupInfo(app = depset(app), deps = depset(deps)),
-    ]
-
-jar_split = rule(
-    implementation = _impl,
-    doc = "Partitions a JVM binary's transitive runtime jars into 'app' " +
-          "(first-party, empty workspace) and 'deps' (third-party) output groups.",
-    attrs = {
-        "binary": attr.label(
-            mandatory = True,
-            providers = [[JavaInfo]],
-            doc = "A jvm binary/library target whose runtime classpath to split.",
-        ),
-    },
-)
@@ -32,13 +32,6 @@ services:
      DO_SPACES_ENDPOINT: "${DO_SPACES_ENDPOINT:-https://sfo3.digitaloceanspaces.com}"
      DO_SPACES_ACCESS_KEY: "${DO_SPACES_ACCESS_KEY:-}"
      DO_SPACES_SECRET_KEY: "${DO_SPACES_SECRET_KEY:-}"
-      EAGLE_HISTORY_BACKEND: "${EAGLE_HISTORY_BACKEND:-postgres}"
-      EAGLE_POSTGRES_HOST: "${EAGLE_POSTGRES_HOST:-}"
-      EAGLE_POSTGRES_PORT: "${EAGLE_POSTGRES_PORT:-}"
-      EAGLE_POSTGRES_DATABASE: "${EAGLE_POSTGRES_DATABASE:-}"
-      EAGLE_POSTGRES_USER: "${EAGLE_POSTGRES_USER:-}"
-      EAGLE_POSTGRES_PASSWORD: "${EAGLE_POSTGRES_PASSWORD:-}"
-      EAGLE_POSTGRES_SSLMODE: "${EAGLE_POSTGRES_SSLMODE:-require}"
      # JWT public key for token validation (auth service handles signing)
      # Reads from /etc/eagle0/keys/public.pem via shared volume
      # Auth token for Shardok on Hetzner (required)
@@ -88,13 +81,6 @@ services:
      DO_SPACES_ENDPOINT: "${DO_SPACES_ENDPOINT:-https://sfo3.digitaloceanspaces.com}"
      DO_SPACES_ACCESS_KEY: "${DO_SPACES_ACCESS_KEY:-}"
      DO_SPACES_SECRET_KEY: "${DO_SPACES_SECRET_KEY:-}"
-      EAGLE_HISTORY_BACKEND: "${EAGLE_HISTORY_BACKEND:-postgres}"
-      EAGLE_POSTGRES_HOST: "${EAGLE_POSTGRES_HOST:-}"
-      EAGLE_POSTGRES_PORT: "${EAGLE_POSTGRES_PORT:-}"
-      EAGLE_POSTGRES_DATABASE: "${EAGLE_POSTGRES_DATABASE:-}"
-      EAGLE_POSTGRES_USER: "${EAGLE_POSTGRES_USER:-}"
-      EAGLE_POSTGRES_PASSWORD: "${EAGLE_POSTGRES_PASSWORD:-}"
-      EAGLE_POSTGRES_SSLMODE: "${EAGLE_POSTGRES_SSLMODE:-require}"
      # JWT public key for token validation (auth service handles signing)
      # Reads from /etc/eagle0/keys/public.pem via shared volume
      SHARDOK_AUTH_TOKEN: "${SHARDOK_AUTH_TOKEN:-}"
@@ -68,7 +68,7 @@ private GameObject GetEffectPrefab(string beastName) {

 ### 4. Wire up in the scene

-In `Assets/Scenes/Eagle.unity`, find the `ProvinceBeastsController` component on the Eagle map object and assign your new prefab to the new field.
+In `Gameplay.unity`, find the `ProvinceBeastsController` component and assign your new prefab to the new field.

 ### 5. Test

@@ -1,149 +0,0 @@
-# Postgres History Migration Plan
-
-## Goal
-
-Evaluate and, if the measurements support it, migrate production game history persistence from local SQLite plus S3 base
-and delta files to a managed Postgres database.
-
-This is not just a mechanical port. The important product requirement is that battle commands, especially Shardok map
-clicks, show results to the player as quickly as possible while still keeping the chance of lost results acceptably low.
-
-## Proposed Direction
-
-Use Postgres as the authoritative production store. Local SQLite should remain useful for development, tests, and rollback
-during rollout, but production should not have two long-lived authoritative stores for the same game.
-
-Start with a small DigitalOcean Managed Postgres instance in the same region and VPC as Eagle. The $15 single-node plan is
-reasonable for early measurement and staging. If the design works, production can move to a more reliable plan when the
-availability requirements justify it.
-
-## Phase 1: Measure First
-
-Before porting the full history implementation, add a benchmark or admin utility that runs from the Eagle deployment
-environment against the candidate Postgres database.
-
-Measure at least:
-
- single action-result append transaction latency
- batch append latency for 25 and 100 action results
- latest snapshot lookup
- replay from a snapshot to a target index
- `since` and stream-update query latency
- Shardok battle result append latency
- p50, p95, p99, and max latency
- async queue lag, if an async persistence prototype is included
-
-Useful initial exit criteria:
-
- Strategic action-result writes should be comfortably below 50 ms p95.
- Shardok synchronous persistence should be below roughly 20-30 ms p95 to stay on the click-to-animation path.
- If Shardok sync writes are above that, use an async Shardok persistence design rather than blocking player feedback.
-
-## Phase 2: Schema Prototype
-
-Add Postgres configuration and a prototype schema that maps the current SQLite history model.
-
-Expected tables:
-
- `games`, for game-level metadata and backend bookkeeping
- `action_results`, for ordered strategic action results
- per-entity decomposition tables currently written by the SQLite dispatchers
- `state_snapshots`, for replay anchors
- Shardok battle result tables, including battle id, sequence number, command/result payloads, and player metadata
- `schema_migrations`, unless the migration mechanism is handled outside the app
-
-Prefer batched inserts and batched reads. The current SQLite implementation can afford some local-file patterns that will
-be too chatty over a network.
-
-## Phase 3: Implement Postgres History
-
-Implement a `FullGameHistory` backend backed by Postgres. Keep the existing SQLite backend available behind a config flag
-during rollout.
-
-Core methods to support:
-
- `count`
- `last`
- `withNewResults`
- `stateAfter`
- `since`
- `sinceDate`
- round-local recent-result queries
- Shardok result writes and reads
-
-The implementation should preserve action ordering and idempotency. If retries are possible, writes should include stable
-keys or sequence ranges so replaying a write request cannot duplicate results.
-
-## Phase 4: Decide Shardok Durability Mode
-
-Shardok latency is the place where this migration can most visibly hurt the player. There are two viable modes.
-
-### Option A: Synchronous Shardok Writes
-
-Use this if measured Postgres latency is low enough. The command path writes results to Postgres before telling the client
-about them.
-
-This is the simplest durability model, but only acceptable if it keeps click-to-animation latency low.
-
-### Option B: Async Shardok Writes
-
-Use this if synchronous Postgres writes add noticeable latency.
-
-The model:
-
- apply battle results to authoritative in-memory game state immediately
- send client updates immediately after the ordered in-memory apply
- enqueue ordered persistence batches in the same order they were applied
- persist batches with `battle_id`, sequence range, payload hash, and payload
- track `lastAppliedSeq` and `lastPersistedSeq`
- retry failed writes until they succeed
- alert when persistence lag exceeds a small threshold
- block or degrade new commands if the queue grows too large
- force a synchronous checkpoint before battle end reconciliation or strategic-layer handoff
- drain the queue during graceful shutdown
-
-This creates a small crash-loss window, but only for rare crashes that happen after the player sees a result and before
-the async write completes. That tradeoff may be better than adding network persistence latency to every tactical command.
-
-## Phase 5: Rollout
-
-Add a backend flag such as:
-
-```text
-EAGLE_HISTORY_BACKEND=sqlite|postgres
-```
-
-Suggested rollout:
-
-1. Run benchmarks from the deployed environment.
-2. Enable Postgres in staging for new games.
-3. Run warmup games and Shardok battle smoke tests.
-4. Enable Postgres for new production games.
-5. Migrate old games either lazily on first open or through an admin-triggered migration job.
-6. Keep SQLite/S3 rollback available until enough new-game production time has passed.
-
-If Postgres is unavailable, commands should not be acknowledged as durable. Either hold/retry them before applying, or
-reject them clearly. Do not show a result to the client while pretending it has already been durably stored unless that is
-an explicit async-durability path with lag monitoring.
-
-## Phase 6: Cleanup
-
-After production confidence:
-
- remove production dependence on S3 base/delta history files
- keep object storage for backups, exports, or emergency archives
- keep SQLite for local development and tests if it remains useful
- document restore, migration, and rollback procedures
-
-## Open Questions
-
- What exact durability guarantee do we want for "the client already saw this result"?
- What Shardok persistence latency is acceptable in real play?
- Should old games migrate lazily or through a bulk/admin process?
- When should the production database move from single-node to HA?
- What should the player experience be if Postgres is reachable but slow?
-
-## Recommended Next PR
-
-The next implementation PR should add a Postgres benchmark/admin utility and configuration plumbing, not the full port.
-That gives us real latency numbers from the Eagle deployment environment before committing to the larger migration.
@@ -136,7 +136,7 @@ Create a tool to programmatically regenerate the Eagle strategic map images with
 | `Assets/Eagle/Materials/*.png` | Replaced province masks |
 | `Assets/Eagle/map_bw_labels.png` | Either: regenerated, or removed if using dynamic labels |
 | `province_map.tsv` | Add centroid columns |
-| `Assets/Scenes/Eagle.unity` | Add ProvinceLabelsController, label prefab instances |
+| `Assets/Gameplay.unity` | Add ProvinceLabelsController, label prefab instances |

 ---

@@ -1,77 +0,0 @@
-# Shardok AI Experiment Tooling Plan
-
-## Summary
-
-Build a reusable workflow for Shardok AI battle baselines and scoring experiments. The goal is to
-turn the current one-off benchmark process into repeatable commands that extract real battles, run
-baseline and candidate configurations, capture detailed per-command traces, and write comparable
-reports.
-
-This work starts with the target-aware archery scoring experiment: attacker `EXPERIMENTAL` against
-defender `STANDARD`.
-
-## Key Changes
-
- Add simulator trace support to `ai_battle_simulator_main`.
-  - New flag: `--trace-jsonl=<path>`.
-  - Emit one JSONL row per AI-selected command.
-  - Include run label, config path, state file, phase, sequence, round, player, side, command type,
-    actor unit, target coords, chosen index, available command count, search depth, commands
-    evaluated, completion reason, forced commands posted afterward, and pre/post side troop totals.
-  - Keep existing summary output unchanged.
-
- Add reusable benchmark tooling under
-  `src/main/cpp/net/eagle0/shardok/ai_battle_simulator/benchmarks/`.
-  - `real_battle_suite.py`: shared helpers for reading `game.db`, materializing battle/state/result
-    blobs, extracting configs, running simulator binaries, parsing summaries, and collecting
-    real-game metadata.
-  - `run_real_battle_experiment.py`: CLI orchestration for baseline/candidate runs.
-  - Outputs: `summary.csv`, `comparison.csv`, `report.md`, generated configs, and per-run trace
-    JSONL files.
-
- Add a clean EXPERIMENTAL switch for target-aware archery value.
-  - Use a new experiment id.
-  - Change only archery-availability scoring.
-  - Replace the flat archery-possible value with a target-sensitive value based on the best enemy
-    unit value.
-  - Do not include the previous fear-value change from experiment `27`.
-
-## Execution Plan
-
- Create a feature branch from fetched `origin/main`.
- Commit this plan document first.
- Implement simulator trace output and reusable benchmark tooling.
- Add the target-aware archery-only EXPERIMENTAL switch.
- Run the new tool first on battle `6248`, then on the all-40 saved-battle suite if local runtime is
-  acceptable.
- Compare baseline attacker `STANDARD` / defender `STANDARD` with candidate attacker
-  `EXPERIMENTAL` / defender `STANDARD`.
- Record the results under the existing benchmark results area and call out battle `6248` in the
-  generated report.
- Push the branch and create a PR after local validation.
-
-## Tests
-
- Build:
-  - `bazel build //src/main/cpp/net/eagle0/shardok/ai_battle_simulator:ai_battle_simulator_main`
-  - `bazel build //src/main/cpp/net/eagle0/shardok/ai_battle_simulator:real_battle_config_extractor`
-  - `bazel build //src/main/cpp/net/eagle0/shardok/ai_battle_simulator:real_battle_smoke_metadata`
- Unit/integration:
-  - `bazel test //src/test/cpp/net/eagle0/shardok/ai_battle_simulator:ai_battle_simulator_test`
-  - Add or extend a simulator test that verifies trace output is created and contains required
-    command fields.
-  - `python3 -m py_compile` on the new benchmark scripts.
- Validation:
-  - Run the new tool on battle `6248`.
-  - Run the new tool on all 40 battles if feasible locally.
-  - Run `bazel test //src/test/cpp/...` before push if local time allows; otherwise open the PR
-    promptly and continue validation while CI starts.
-
-## Assumptions
-
- "Experiment 3" means target-aware archery value from the previous recommendation list.
- The first experiment compares attacker-only candidate behavior against unchanged defender
-  `STANDARD`.
- JSONL is the trace format because command diagnostics will evolve over time.
- Python benchmark scripts follow the existing executable-script pattern rather than adding Python
-  Bazel targets.
@@ -1,82 +0,0 @@
-# Shardok AI Scoring Benchmarks
-
-The current STANDARD scoring experiment results are recorded in
-`src/main/cpp/net/eagle0/shardok/ai_battle_simulator/benchmarks/experimental_results/`.
-They use 40 saved Shardok battles extracted from `game_389ffb1901903118.zip`.
-
-The corrected all-40 run uses:
-
- `max_rounds = 31`
- `random_seed = 1`
- iterative-deepening timeout propagation fix from PR #6923
- AI setup placement
- forced single-option commands
- defender scoring fixed to `STANDARD` for every row
- archery/start-fire capability derived like production setup
- per-unit flee capability preserved by real-battle extraction
- zero-troop combat heroes preserved as live units
- `NO_PROFESSION` treated as a valid hero profession state, not as "no hero"
-
-Shardok siege battles do not have a draw result. If the attacker has not won after 31 rounds, the
-defender wins. The simulator records those cutoffs as `max_rounds_reached` defender wins.
-
-All 40 saved battle payloads now produce valid simulator configs. A battle side can have zero
-remaining troops as long as it has a hero; this is common for `NO_PROFESSION` heroes with no
-commanded battalion.
-
-An all-40 real-vs-sim comparison records whether the saved-game attacker/defender were human or AI
-players. Factions `3` and `4` are human players; all other factions in this save are AI players.
-The comparison uses broad winner side from the saved Shardok result payload, because exact replay
-is not available from the compact result rows. Real end-of-battle troop totals come from the
-all-factions Shardok `ActionResultView` stream's final `changed_player_totals`.
-
-An AI-vs-AI smoke check also reruns a subset where neither side used faction `3` or `4`.
-
-## Current Results
-
-These artifacts are the current baseline for future scoring comparisons.
-
-| Dataset | Valid battles | Invalid rows | Attacker wins | Defender wins | Notes |
-| --- | ---: | ---: | ---: | ---: | --- |
-| All-40 timeout-fix STANDARD baseline | 40 | 0 | 18 | 22 | `standard_scoring_all40_post_bugfix_baseline.*`; regenerated after fixing iterative-deepening timeout score propagation |
-| All-40 real saved game | 40 known | 0 unknown | 19 | 21 | `standard_scoring_all40_real_vs_sim.*`; sim matches 39/40 known winners |
-| AI-vs-AI smoke | 31 | 0 | 10 | 21 | 31/31 winner-side matches |
-
-The remaining real-vs-sim winner-side miss is `11754` (real attacker, sim defender), which had an
-AI attacker but a human defender in the saved game. The real attacker won with `2119` troops
-remaining; the STANDARD-vs-STANDARD simulation replaces the human defender with AI and the defender
-wins with `1000` troops remaining, so this row is not a comparable AI-vs-AI smoke miss. Battle
-`6248`, previously the remaining AI-vs-AI miss, now matches as an attacker win after expected-impact
-archery scoring was promoted to STANDARD.
-
-The closest six valid baseline rows are used for the scalar and behavior experiment batches. The
-single-run harness still has some nondeterminism even with `random_seed = 1`, so near-margin changes
-should be repeated before being promoted.
-
-## Experiment Tooling
-
-`src/main/cpp/net/eagle0/shardok/ai_battle_simulator/benchmarks/run_real_battle_experiment.py`
-is the reusable real-battle experiment runner. It can read a saved-game zip or extracted `game.db`,
-generate baseline/candidate simulator configs, run both sides, and write `summary.csv`,
-`comparison.csv`, `report.md`, generated configs, and per-command trace JSONL files.
-
-The repository keeps the current STANDARD baseline artifacts, but omits generated configs/traces
-for abandoned experiment batches. Regenerate those locally with the runner when a new comparison
-needs command-level traces.
-
-`src/main/cpp/net/eagle0/shardok/ai_battle_simulator/benchmarks/run_standard_baselines.py`
-regenerates the tracked STANDARD-vs-STANDARD baseline artifacts from a saved-game zip or extracted
-`game.db`, including the all-40 baseline, real-vs-sim comparison, and AI-vs-AI smoke files.
-
-The simulator supports `--trace-jsonl=<path>` and `--run-label=<label>` for detailed command traces.
-Each row records the selected command, search depth, forced command count, and pre/post side troop
-totals.
-
-Expected-impact archery replaced the old flat archery-position value in STANDARD. A unit that can
-shoot now gets half of the expected volley impact; a unit that is merely positioned for a future
-volley gets one quarter. Promoting that behavior flips `6248` to match the real AI-vs-AI attacker
-win without introducing an AI-vs-AI winner-side miss in the current saved-battle baseline.
-
-The latest baseline refresh was generated after the iterative-deepening evaluator stopped letting
-timeout fallback scores compete with completed command evaluations. Future STANDARD-vs-EXPERIMENTAL
-comparisons should compare against these checked-in timeout-fix baseline artifacts.
@@ -6,16 +6,6 @@ Be able to support a small (10-50 user) private alpha, including with strangers.

 Larger set of goals in [The Big Eagle TODO](https://docs.google.com/document/d/17RTt3-4Wl2AAVMRLodaC3a84E4de6xuQTWBPvCRM484/edit?pli=1&tab=t.0), but most of that is not necessary for MVP.

-## Current priority order
-
-1. Finish tutorial & first-session onboarding, especially the clear first-session goal and guided first scenario.
-2. Fix the basic Shardok AI issues that can make tactical battles feel broken or unfair.
-3. Publish a known issues doc so alpha testers do not repeatedly report the same rough edges.
-4. Verify the existing game-ended flow, then mark the win condition done if "all other factions defeated" works end-to-end.
-5. Add mid-game progression events where the King recognizes the player as they gain power.
-6. Add account-linking when provider switching or account recovery becomes important for alpha support.
-7. Leave terms of service and Windows code signing deferred until they matter for public release or broader distribution.
-
 ## Required

 ### Gameplay Productionization
@@ -27,7 +17,7 @@ Larger set of goals in [The Big Eagle TODO](https://docs.google.com/document/d/1
 - [x] ~~Fix long disconnects on deployments~~
 - [x] Fix the Mac installer
 - [x] ~~Still not reconnecting after deployments~~
- [x] Notify about client updates, button to come directly back
+- [ ] Notify about client updates, button to come directly back
 - [x] Generatedtext healing
 - [x] Kill outstanding shardok requests when game is deleted

@@ -1,335 +0,0 @@
-# SQLite Action Result Decomposition
-
-Follow-up to [SQLITE_HISTORY_DESIGN.md](SQLITE_HISTORY_DESIGN.md). This document plans the second wave of the SQLite migration: decomposing the `ActionResult` proto blob in `action_results.payload` into native relational tables, deleting the proto representation entirely for everything in scope.
-
-## Status — option B adopted (updated 2026-05-17)
-
-**This section supersedes the original plan below. Where the body conflicts with this section, this section wins.** The full-relational design is kept below for historical context and as the migration target if a structured-query need on deep aggregates ever materializes.
-
-The original plan (decompose every field into relational tables; "decision locked") was implemented for `ChangedProvince` first — ~37 child tables, ~1500 lines (PR #6721) — and then **rejected on cost/benefit**. We adopted **option B**: apply the "stopping rule" (originally carved out only for `Quest`) *consistently*. Shipped and merged as Phase 4.5b in **PR #6725**:
-
- Keep the per-entity **top-level table with scalar / entity-reference columns + indexes** (the 4.5a tables). That is the part that earns its keep — "which actions changed province/hero/faction X, by how much" is one-line SQL.
- Store the **rest of the aggregate as one `<entity>_proto BLOB` column** on that row: lossless round-trip / replay, no child tables for the deep nesting.
- Scalar columns are pure query denormalization; **the proto blob is authoritative**. A later fully-relational pass (if ever needed) is a pure read-blob → write-columns migration — which is why the blob must survive any future proto deletion.
-
-Rationale: the deep army/unit/hero/claimant nesting is never analytics-queried — the exact property that justified blobbing `Quest`. Exploding it into ~37 tables was code volume without query value. The blob also means deep-aggregate field changes need **no SQL migration** (only promoted scalar columns do), which is the right shape for the volatile parts of the model. This rescinds "Decisions locked #1" and revises principle 7, the Scope, the Sequencing table, and "Deleting the proto" (all updated in place below).
-
-## Refinement — "refined option B" for shallow entities (updated 2026-05-19)
-
-**This section further supersedes the body and the Status section above where they conflict; it wins.** It records the pattern actually shipped for `ChangedHero` in Phase 4.5c.1 (#6727), which is a refinement of option B, not the whole-entity-blob shape the body still describes for that entity.
-
-Option B as originally stated keeps **one `<entity>_proto BLOB`** per row for the entire non-scalar residual. For a *shallow* entity that is more blob than it needs. Refined option B decomposes the residual into three tiers instead of one blob:
-
- **Tier 1 — scalar / entity-reference columns + indexes.** The queryable surface (deltas, touched entities, enum ordinals). Always present; unchanged from 4.5a.
- **Tier 2 — typed columns for sealed value-oneofs.** A sealed choice over simple values (e.g. `ChangedHeroC`'s `loyalty` / `vigor` `StatChange`, subtypes `StatDelta`/`StatAbsolute`/`StatNoChange`) becomes a **pair of nullable columns** (`loyalty_delta` / `loyalty_absolute`, …). The *column name is the discriminator*; at most one of each pair is non-NULL; both NULL ⇒ the no-change case. No `kind` column.
- **Tier 3 — per-row leaf-proto blobs in a child table.** A repeated/optional nested aggregate becomes a child table keyed by `(action_seq, entity_id, n)` storing the **leaf** proto's bytes one row per element (e.g. `action_changed_heroes_new_backstory_events` holding `EventForHeroBackstory` blobs) — *not* a whole-entity blob.
-
-**Consequence (changes Decisions #1–#3 and 4.5i):** an entity fully covered by Tiers 1–3 has **no whole-entity blob**. It round-trips losslessly from columns + per-leaf blobs, so its per-entity proto and converter (`changed_hero.proto` / `ChangedHeroConverter`) become genuinely dead code, deletable in 4.5i. Only the **leaf** protos used as Tier-3 blob payloads (`event_for_hero_backstory.proto`) are retained as a wire format. The whole-entity `<entity>_proto BLOB` is now reserved for **deep** entities where Tiers 2–3 would be more work than value (`ChangedProvince`, the "new entity" snapshots).
-
-Per-entity decision (supersedes Decisions #2): classify each entity as **all-scalar** (Tier 1 only, no blob — e.g. likely `ChangedFaction`), **refined** (Tiers 1–3, no whole-entity blob, proto deletable — `ChangedHero`), or **deep** (Tier 1 + whole-entity `<entity>_proto BLOB`, proto retained — `ChangedProvince`, new-entity snapshots). Default to **refined** for shallow entities; fall back to **deep** only when the residual is too varied to be worth Tiers 2–3. Wherever the body or Status says "the proto blob is authoritative," for a *refined* entity read "the columns + per-leaf-proto blobs are authoritative."
-
-## Why
-
-The first wave (Phases 1–4) stores each action as a single proto blob in `action_results.payload`. That blob is opaque to SQL — any query about action contents (changed provinces, heroes affected, etc.) requires deserializing every row. The `action_result_type`, `round_id`, and date columns we denormalized are the only handles SQL has into actions.
-
-Two observations make full decomposition more attractive than the original design admitted:
-
-1. **The proto exists for storage only.** 92 Scala files reference `ChangedProvince`, but exactly one imports the proto type — `ChangedProvinceConverter.scala`. Everything else (action handlers, AI, utils) operates on `ChangedProvinceC`. The proto is **not** in any client-facing RPC contract — it lives under `eagle.internal`. Same is true for `ChangedHero`, `ChangedFaction`, `ChangedBattalion`, `Notification`, `GeneratedTextRequest`, and `ActionResult` itself. Replacing proto serialization with SQL writes does **not** lose forward/backward compatibility, because there are no external consumers depending on the proto wire format.
-
-2. **`ActionResultC` has no polymorphism at the top level.** It's a 34-field flat struct — `Option`s and `Vector`s, but no `sealed trait` discriminator at the action level. Every action has the same shape; the `actionResultType` enum tells you *which fields* are populated. This means the entire structure is relationalizable without per-action-type tables.
-
-The work is bounded, the proto becomes dead code at the end of it, and it has to happen before alpha — once we have users, every schema change costs a real migration. Doing it pre-alpha lets us nuke save dirs as part of deployment.
-
-## Scope
-
-**In:** `action_results.payload` (the whole-`ActionResult` blob) is gone by the end. Per-entity scalar / entity-reference columns make the *queryable* surface (deltas, touched entities) native SQL. **Under refined option B (see Status):** `internal/action_result.proto` and *deep* entities' sub-message protos (`changed_province.proto`, …) are retained as blob payloads; *refined* entities' protos (`changed_hero.proto`) are decomposed away and deleted, with only their Tier-3 leaf protos (`event_for_hero_backstory.proto`) kept. (The original plan deleted all of them, then a draft of option B kept all of them; neither holds — it's now per-entity.)
-
-**Out:**
- `state_snapshots.payload` — stays as proto-blob `GameState`. These are caches for replay, not authoritative data; querying inside a snapshot adds nothing over querying the action deltas that produced it. Cost/benefit doesn't justify decomposing them.
- Shardok tables — already designed in Phase 3, no changes here. The shardok protos (`ShardokActionResult`, etc.) are owned by the Shardok subsystem and are part of an actual RPC contract; they stay.
- Client text — already SQLite-backed (`SqliteClientTextStore`).
-
-## Principles
-
-1. **One table per non-trivial aggregate type.** Every `Vector[X]` or `Option[X]` where `X` is a Scala case class with multiple fields gets its own table.
-2. **Vectors of primitives → join tables.** `Vector[Int]` becomes a two-column table: `(parent_key, value)`. No JSON arrays, no comma-separated strings — those defeat the point.
-3. **Optional aggregates → presence by row existence.** `Option[X]` for aggregate `X` is a row that may or may not exist with the matching parent key. The natural primary key on `(action_seq)` enforces 0-or-1-row semantics for `Option[X]` aggregates and `(action_seq, n)` for vectors.
-4. **Sealed traits get a discriminator column.** When a Scala sealed trait has a few subtypes with diverging fields, prefer one wide table with `kind TEXT NOT NULL` + nullable per-subtype columns over per-subtype tables. When subtypes diverge significantly (>5 unrelated fields each), use per-subtype tables. Decide per case.
-5. **`ON DELETE CASCADE` everywhere.** Every child table has `FOREIGN KEY (action_seq) REFERENCES action_results(action_seq) ON DELETE CASCADE`. `truncateTo` becomes a single `DELETE FROM action_results WHERE action_seq >= ?` and the cascades do the rest.
-6. **Schema migrations are now real.** Once we have a relational schema, evolving it means SQL migrations. We introduce a real migration runner (see "Schema evolution" below). The `metadata` table's `schema_version` becomes load-bearing.
-7. **(Revised under option B — see Status.) The proto blob is the retained authoritative store for each entity's deep aggregates; scalar / entity-reference columns are denormalized projections of it.** The original principle assumed full decomposition and proto deletion. Instead each entity keeps one proto-blob source of truth and denormalizes only the queryable columns from the same write — so there is no dual *relational* representation to keep in sync.
-
-## Top-level: `action_results`
-
-The `payload BLOB` column goes away. Scalar fields become columns; aggregate fields move to related tables.
-
-```sql
-CREATE TABLE action_results (
-    action_seq                              INTEGER PRIMARY KEY,
-    action_result_type                      INTEGER NOT NULL,
-    -- existing scalar denormalizations
-    round_id                                INTEGER NOT NULL,
-    date_year                               INTEGER,
-    date_month                              INTEGER,
-    -- new scalar columns (formerly inside payload)
-    acting_hero_id                          INTEGER,
-    acting_faction_id                       INTEGER,
-    province_id                             INTEGER,
-    province_id_acted                       INTEGER,
-    new_round_phase                         INTEGER,
-    new_round_id                            INTEGER,
-    last_command_type_for_acting_province   INTEGER,
-    resolved_battle                         TEXT,
-    new_victor_faction_id                   INTEGER,
-    game_ended                              INTEGER,  -- 0/1/NULL
-    new_random_seed                         INTEGER,
-    new_game_type                           INTEGER
-);
-
-- Vectors of bare IDs become join tables:
-CREATE TABLE action_destroyed_battalion_ids (
-    action_seq    INTEGER NOT NULL,
-    battalion_id  INTEGER NOT NULL,
-    PRIMARY KEY (action_seq, battalion_id),
-    FOREIGN KEY (action_seq) REFERENCES action_results(action_seq) ON DELETE CASCADE
-);
-
-CREATE TABLE action_removed_hero_ids (
-    action_seq INTEGER NOT NULL,
-    hero_id    INTEGER NOT NULL,
-    PRIMARY KEY (action_seq, hero_id),
-    FOREIGN KEY (action_seq) REFERENCES action_results(action_seq) ON DELETE CASCADE
-);
-
-CREATE TABLE action_removed_faction_ids (...);
-CREATE TABLE action_affected_faction_ids (...);
-
-- Singleton optional aggregates: a row exists if and only if Option is Some.
-CREATE TABLE action_new_battles (
-    action_seq INTEGER PRIMARY KEY,
-    -- ShardokBattle fields here, columns or sub-tables as needed
-    FOREIGN KEY (action_seq) REFERENCES action_results(action_seq) ON DELETE CASCADE
-);
-
-CREATE TABLE action_new_chronicle_entries (action_seq INTEGER PRIMARY KEY, ...);
-CREATE TABLE action_new_eagle_map_info (action_seq INTEGER PRIMARY KEY, ...);
-```
-
-Each `Vector[Aggregate]` field on `ActionResultC` gets its own table (covered below).
-
-### Indexes
-
-We keep the existing indexes (`round_id`, `(date_year, date_month)`) and add a few that obvious queries need:
-
- `idx_action_results_acting_hero_id`
- `idx_action_results_acting_faction_id`
- `idx_action_results_province_id`
-
-Plus indexes on the per-entity join tables on the entity-id side, so cross-game queries like "all actions touching hero X" become single-index lookups:
-
- `idx_action_changed_provinces_province_id`
- `idx_action_changed_heroes_hero_id`
- `idx_action_changed_faction_faction_id`
- `idx_action_changed_battalions_battalion_id`
-
-## Changed-entity tables (four parallel structures)
-
-### `action_changed_provinces`
-
-`ChangedProvinceC` has ~50 fields: scalar deltas, nested aggregates, ID vectors. The scalar fields become columns on `action_changed_provinces`; aggregate fields become child tables.
-
-```sql
-CREATE TABLE action_changed_provinces (
-    action_seq                            INTEGER NOT NULL,
-    province_id                           INTEGER NOT NULL,
-    -- resource changes
-    gold_delta                            INTEGER,
-    food_delta                            INTEGER,
-    -- stat changes
-    new_price_index                       REAL,
-    economy_delta                         REAL,
-    agriculture_delta                     REAL,
-    infrastructure_delta                  REAL,
-    economy_devastation_delta             REAL,
-    agriculture_devastation_delta         REAL,
-    infrastructure_devastation_delta      REAL,
-    support_delta                         REAL,
-    -- round properties
-    set_has_acted                         INTEGER,  -- 0/1/NULL
-    set_ruler_is_traveling                INTEGER,
-    -- ruling faction changes
-    clear_ruling_faction_id               INTEGER NOT NULL DEFAULT 0,
-    new_ruling_faction_id                 INTEGER,
-    -- misc scalar fields
-    new_locked_improvement_kind           TEXT,     -- 'none', 'new'
-    new_locked_improvement_value          INTEGER,  -- ImprovementType if kind='new'
-    new_province_orders                   INTEGER,  -- ProvinceOrderType enum
-    clear_defending_army                  INTEGER NOT NULL DEFAULT 0,
-    cleared_pending_conquest_info         INTEGER NOT NULL DEFAULT 0,
-    removed_deferred_change_index         INTEGER,
-    PRIMARY KEY (action_seq, province_id),
-    FOREIGN KEY (action_seq) REFERENCES action_results(action_seq) ON DELETE CASCADE
-);
-
-CREATE INDEX idx_action_changed_provinces_province_id
-    ON action_changed_provinces(province_id);
-```
-
-The vector and optional-aggregate fields of `ChangedProvinceC` become per-aggregate child tables, all keyed by `(action_seq, province_id)`:
-
- `action_changed_provinces_new_unaffiliated_heroes(action_seq, province_id, n, ...)` — one row per added hero, with `UnaffiliatedHero` fields as columns
- `action_changed_provinces_changed_unaffiliated_heroes(...)` — same shape
- `action_changed_provinces_removed_unaffiliated_hero_ids(action_seq, province_id, hero_id)` — join table for ID vector
- `action_changed_provinces_new_captured_heroes(...)` — `CapturedHero` fields
- `action_changed_provinces_recruitment_attempted_captured_hero_ids(...)` — join table
- `action_changed_provinces_removed_captured_hero_ids(...)` — join table
- `action_changed_provinces_new_ruling_faction_hero_ids(...)` — join table
- `action_changed_provinces_removed_ruling_faction_hero_ids(...)` — join table
- `action_changed_provinces_new_battalion_ids(...)` — join table
- `action_changed_provinces_removed_battalion_ids(...)` — join table
- `action_changed_provinces_new_incoming_armies(...)` — `MovingArmy` decomposition
- `action_changed_provinces_removed_incoming_army_ids(...)` — join table
- `action_changed_provinces_new_withdrawing_armies(...)` — `MovingArmy` decomposition
- `action_changed_provinces_removed_withdrawing_army_ids(...)` — join table
- `action_changed_provinces_new_hostile_armies(...)` — `HostileArmyGroup` decomposition
- `action_changed_provinces_removed_hostile_army_faction_ids(...)` — join table
- `action_changed_provinces_hostile_army_status_changes(action_seq, province_id, faction_id, new_status)` — small flat aggregate
- `action_changed_provinces_new_defending_army` — `Option[Army]` (singleton row if present)
- `action_changed_provinces_new_incoming_shipments(...)` — `MovingSupplies` decomposition
- `action_changed_provinces_removed_incoming_shipment_ids(...)` — join table
- `action_changed_provinces_new_incoming_end_turn_actions(...)` — `IncomingEndTurnAction` decomposition
- `action_changed_provinces_removed_incoming_end_turn_actions(...)` — same shape (these are values, not IDs)
- `action_changed_provinces_new_battle_revelations(...)` — `BattleRevelation` decomposition
- `action_changed_provinces_removed_battle_revelations(...)` — same shape
- `action_changed_provinces_new_pending_conquest_info` — `Option[PendingConquestInfo]` (singleton row)
- `action_changed_provinces_new_province_events` — `Vector[ProvinceEvent]`
- `action_changed_provinces_new_deferred_change` — `Option[DeferredChange]` (singleton, with discriminator since `DeferredChangeT` is sealed)
-
-That's a lot of tables — ~25 just under `ChangedProvince`. The schema is mechanical once the principles are set; what makes it tractable is that every table follows the same shape: parent FK, `n` (vector index) where ordering matters, the aggregate's fields as columns, child aggregates as their own tables.
-
-### `action_changed_heroes`, `action_changed_factions`, `action_changed_battalions`
-
-Same pattern. Each gets its top-level table keyed by `(action_seq, entity_id)`, plus child tables for nested aggregates and ID vectors. Indexes on the entity id (`hero_id`, `faction_id`, `battalion_id`) so cross-game queries are fast.
-
-`ChangedHeroC` is a *refined* entity (see "Refinement" in Status): no whole-entity blob. Its sealed `StatChange` (`StatDelta`/`StatAbsolute`/`StatNoChange`) `loyalty` and `vigor` oneofs are Tier-2 column pairs (`loyalty_delta`/`loyalty_absolute`, `vigor_delta`/`vigor_absolute`; column name discriminates, both-NULL ⇒ `StatNoChange`). Its `newBackstoryEvents: Vector[EventForHeroBackstory]` is Tier-3: per-event `EventForHeroBackstory` blobs, one row each, in `action_changed_heroes_new_backstory_events`. Columns + per-event blobs round-trip losslessly, so `changed_hero.proto`/`ChangedHeroConverter` become deletable (4.5i) and only `event_for_hero_backstory.proto` is retained as a blob format.
-
-## "New" entity tables
-
-`ActionResultC.newProvinces: Vector[ProvinceT]`, `newHeroes`, `newFactions`, `newBattalions`. These are *full snapshots* of new entities at creation.
-
-These are bigger and recursive: `ProvinceT` itself has fields including its armies, heroes, battalions, events, etc. Decomposing fully duplicates the structure already covered by the existing `state_snapshots` blob (which we're explicitly keeping). The decomposition still needs to happen because we're dropping the proto entirely — and these vectors live in `ActionResultC`, which we're decomposing.
-
-**Decision**: go fully relational here too. Each "new entity" gets its top-level table (`action_new_provinces` etc.) keyed by `(action_seq, entity_id)`, with the same per-aggregate-field decomposition pattern as the changed-entity tables. Where a "new entity" type contains its own nested aggregates, those aggregates either share child tables with the `changed_*` family (preferred where the field shape is identical, e.g., `MovingArmy`, `UnaffiliatedHero`) or get their own (where they diverge).
-
-This is the most expensive single chunk of the decomposition — explicitly called out so we know what we're committing to in PR 4.5f.
-
-## Notifications and generated text requests
-
-```sql
-CREATE TABLE action_new_notifications (action_seq INTEGER NOT NULL, n INTEGER NOT NULL, ...);
-CREATE TABLE action_removed_notifications (action_seq INTEGER NOT NULL, n INTEGER NOT NULL, ...);
-CREATE TABLE action_new_generated_text_requests (...);
-CREATE TABLE action_client_text_visibility_extensions (...);
-```
-
-`NotificationT` and `GeneratedTextRequestT` are both sealed traits. Use the discriminator-column pattern: `kind TEXT NOT NULL` plus nullable per-subtype columns, unless a subtype's fields are too divergent to share a row.
-
-## Schema evolution
-
-We can't punt on this any more. With this many tables, the next field addition can't be "edit the proto." It needs a real migration mechanism.
-
-Conventions:
- `metadata['schema_version']` is a small integer, currently `1`.
- A `Migrations` object holds an ordered `Vector[(Int, Connection => Unit)]` of migration steps.
- On `SqliteHistory.loaded`, the schema version is read; any pending migrations are applied in a single transaction; the version is bumped.
- New games start at the latest version.
- Migration steps are append-only — never edit an existing step.
-
-This is the standard Rails / Flyway / etc. pattern, scaled down to one file.
-
-## Read path: reconstructing `ActionResultC`
-
-The current pattern:
-```scala
-val proto = ActionResult.parseFrom(payloadBlob)
-val scala = ActionResultProtoConverter.fromProto(proto)
-```
-
-The new pattern:
-```scala
-val scala = ActionResultDbReader.read(connection, actionSeq)  // assembles from action_results + child tables
-```
-
-`ActionResultDbReader.read` is one query per child table (or a single multi-result query with carefully ordered joins, though that has aggregation pitfalls). Per-table queries are simpler to write and SQLite query planning handles them well.
-
-For bulk reads (`since`, `all`, `sinceDate`, etc.), the existing `replayFrom` helper iterates action seqs in order, and currently parses one proto per action. The new equivalent runs one batched query per child table for the relevant action_seq range, materializing rows into a map keyed by action_seq, then walks the range emitting `ActionResultC` instances assembled from the maps. This is more code than the proto version but doesn't change the algorithmic shape.
-
-## Write path
-
-`withNewResults` becomes:
-1. INSERT into `action_results` (scalar fields).
-2. For each non-empty aggregate: INSERT N rows into the corresponding child table.
-
-All inside a single transaction. Per-action insert count is bounded by what the action does — most actions touch 0–3 provinces, 0–5 heroes, 0–10 changed entities total, plus 0–2 notifications. So the typical insert is ~5–20 rows. With WAL and a single per-batch transaction, this is cheap.
-
-## Replay performance
-
-Today's `replayFrom` parses one proto per action. Decomposed, each action requires per-child-table queries. For a 25-action replay between snapshots, that's still small absolute work — maybe 100–500 rows fetched from a handful of indexed tables. Should be faster than proto parsing once the JIT warms up, but won't matter either way at typical sizes.
-
-## Deleting the proto
-
-> **Superseded by refined option B (see Status).** Whole-`ActionResult` `payload` is removed in 4.5h. After that it depends on per-entity classification: *refined* entities' per-entity sub-message protos + converters (e.g. `changed_hero.proto`) **do** become dead code and are deleted in 4.5i; *deep* entities' whole-entity blob protos, the leaf protos used as Tier-3 blob payloads (e.g. `event_for_hero_backstory.proto`), and `Quest` are *retained*. The original full-deletion plan below is kept for historical context.
-
-Once `SqliteHistory` writes and reads exclusively from the new tables:
-
-1. Delete `ActionResultProtoConverter.scala`, the `Changed{Province,Hero,Faction,Battalion}Converter.scala` files, the `Notification`/`GeneratedTextRequest`/etc. converters.
-2. Delete the proto definitions: `action_result.proto`, `changed_province.proto`, `changed_hero.proto`, `changed_faction.proto`, `notification.proto`, `generated_text_request.proto`, anything else in this scope.
-3. Delete the corresponding `_scala_proto` Bazel targets.
-4. Gazelle should drop the unused deps from downstream targets.
-
-This is satisfying but should land *last*, after everything's stable. The proto files are dead code at that point — keeping them around for a PR or two while we verify migration is fine.
-
-## Sequencing
-
-**Phase 4** (cutover) lands as planned: `GamesManager` swaps to `SqliteHistory`, save dirs get nuked, action data is the single proto-blob `payload`. Days of work, validates the lifecycle integration without conflating with schema design.
-
-**Phase 4.5** (this design) starts after Phase 4 is verified stable. Subdivided into small PRs. **Re-planned for option B (see Status):** each entity gets the 4.5a scalar/entity-ref columns + a `<entity>_proto BLOB`; no child-table explosion. Each decomposition PR splits into `.1` (schema + backfill over existing `payload`) and `.2` (live `withNewResults` write path + read path), per the precedent set by 4.5b.
-
-| PR | Work | State |
-|---|---|---|
-| 4.5a | Schema-migration scaffolding + four changed-entity top-level tables (scalar columns) | ✅ merged |
-| 4.5b | `ChangedProvince`: scalar columns + `changed_province_proto BLOB` + v3 backfill (option B, #6725) | ✅ merged |
-| 4.5c.1 | `ChangedHero` (**refined option B**, see Status): 4.5a scalars + loyalty/vigor Tier-2 column pairs + `newBackstoryEvents` as per-event `EventForHeroBackstory` Tier-3 blobs in `action_changed_heroes_new_backstory_events`; **no whole-entity blob** — `changed_hero.proto`/`ChangedHeroConverter` become deletable, `event_for_hero_backstory.proto` retained. Schema v4 + backfill over existing `payload` (#6727) | ✅ merged |
-| 4.5c.2 | `ChangedHero`: live `withNewResults` write path + read path | planned |
-| 4.5d | `ChangedFaction`: same pattern (mostly scalar already; small/empty blob — see Decisions #2) | planned |
-| 4.5e | `ChangedBattalion`: `changed_battalion_proto BLOB` (sole field is the contained `BattalionT`) + battalion_id index from 4.5a | planned |
-| 4.5f | "New entity" vectors (`newProvinces`/`newHeroes`/`newFactions`/`newBattalions`): one table per type keyed by `(action_seq, entity_id)` + a couple of queryable handles + `<entity>_proto BLOB`. Biggest win for B — these are the deepest snapshots | planned |
-| 4.5g | Notifications + generated text requests: `(action_seq, n, kind, <entity>_proto BLOB)` — `kind` discriminator queryable, payload blobbed (quest-hybrid shape) | planned |
-| 4.5h | Top-level `ActionResultC` scalar/entity-ref columns on `action_results` (acting_hero_id, acting_faction_id, province_id, …) + **drop `payload BLOB`**. Read path = scalar columns + per-entity proto blobs (close to today's proto parse; simpler than the original per-table-join plan) | planned |
-| 4.5i | Delete proto files/targets that became dead code. *Refined* entities' per-entity protos + converters (e.g. `changed_hero.proto`/`ChangedHeroConverter`) **are deleted** here. *Deep* entities' protos + leaf Tier-3 protos (e.g. `event_for_hero_backstory.proto`) + `Quest` are **retained** as blob payloads. No longer near-nothing — scope scales with how many entities went *refined* | planned |
-
-Each PR nukes save dirs as part of its rollout (still pre-alpha). Each PR is independently reviewable and revertable. Under option B the per-PR cost is far lower than the original estimate (~110-line writer per entity, no child-table schema), so the remaining sequence is roughly **1 week**, not 3–4.
-
-After 4.5: Phase 5 (idle-game eviction) and Phase 6 (final cleanup) proceed as the parent design doc lays out.
-
-## What we gain at the end
-
- **Single source of truth.** Action data lives in tables; no proto representation for the things in scope.
- **Native SQL analytics.** Cross-game queries like "actions touching province X" or "all riots in year Y" are one-line SELECTs. No backfill code needed.
- **Inspectable saves.** `sqlite3 game.db` shows readable data. Debugging is normal SQL work, not "parse this binary blob."
- **Cleaner replay path.** No proto parsing during replay. The `ActionResultProtoConverter` (currently called per replayed action in `replayFrom`) goes away entirely.
- **Less code.** ~7 proto files deleted; their generated Scala targets deleted; their converter classes deleted; the dual proto/Scala representation collapses to one.
- **Schema migrations as a first-class concept.** Future field additions are SQL migrations, which is the right shape for a system that owns its storage.
-
-## What we lose
-
- **The "edit one proto" workflow** for adding fields. Adding `newFooDelta: Double` to `ChangedProvinceC` is now a Scala edit + SQL migration step + converter update. More steps. Pre-alpha this is cheap; post-alpha (when migrations apply to real data) it's a careful PR but still bounded.
- **A safety net for unrecognized fields.** Proto silently keeps unknown fields on parse; SQL doesn't. We accept this because we control both ends.
-
-## Decisions locked
-
-1. **Refined option B is the pattern (see "Refinement" in Status; rescinds both "fully relational, no blob fallback" and the original whole-entity-blob-per-entity rule).** Per entity, three tiers: scalar/entity-ref columns (Tier 1) + typed columns for sealed value-oneofs (Tier 2) + per-row leaf-proto blobs in child tables (Tier 3). A whole-entity `<entity>_proto BLOB` is used **only for deep entities** where Tiers 2–3 cost more than they're worth. No child-table explosion for deep aggregates.
-2. **Classify each entity: all-scalar / refined / deep (supersedes the old "is a blob needed").** *All-scalar* → Tier 1 only, no blob (likely `ChangedFaction`). *Refined* → Tiers 1–3, no whole-entity blob, per-entity proto deletable (`ChangedHero`). *Deep* → Tier 1 + whole-entity `<entity>_proto BLOB`, proto retained (`ChangedProvince`, new-entity snapshots). Default to *refined* for shallow entities; fall back to *deep* only when the residual is too varied for Tiers 2–3.
-3. **What's retained vs. deleted depends on the classification.** `action_results.payload` (the whole-`ActionResult` blob) still goes away in 4.5h. *Refined* entities' per-entity sub-message protos + converters become dead code, deleted in 4.5i; only the **leaf** protos they use as Tier-3 blob payloads (e.g. `event_for_hero_backstory.proto`) are retained. *Deep* entities' protos + `Quest` stay as whole-entity blob payloads. So 4.5i is no longer near-nothing.
-4. **Sub-phase ordering** as in the Sequencing table; each decomposition PR splits into `.1` (schema + backfill) and `.2` (live write + read).
-5. **Phase 4 landed separately, before 4.5a** (historical, unchanged). Cutover validated the lifecycle integration with the simpler schema before decomposition began.
@@ -1,291 +0,0 @@
-# SQLite Game History Design
-
-## Why
-
-The current persistence layer is file-based: action results are chunked into `.e0a` files, individual results spill to `.e0i` files for crash recovery, state snapshots live in `.e0s` files, and a `directory.e0i` index tracks chunks. This shape has produced three workarounds in recent history:
-
-1. A `games.e0es` cache (PR #6706) to avoid re-reading the running-games list.
-2. Dual-storage in `PersistedActionResult` (proto bytes + parsed proto) to amortize serialization across save flushes.
-3. Lazy `gameState` (PR #6692) to avoid eager proto conversion on every replayed step, after PR #6691 fixed the `stateAfter` OOM that materialized ~4500 `PersistedActionResult` instances during a stale-LLM-replay path.
-
-Each is a workaround for the same root cause: file-based random access is expensive, so we cache aggressively, and the cache layers are fragile (one misplaced `.map(_.gameState)` reintroduces the OOM).
-
-The alternative — SQLite as a per-game container, with proto bytes stored as BLOB columns indexed by action sequence number — gives us random access by construction, removes the need for all three workarounds, and makes idle-game eviction cheap (rehydrate from the DB instead of reparsing files). The pattern is already in production here: `SqliteClientTextStore` does exactly this for the client text cache. We extend the same pattern to game history.
-
-## Goal & non-goals
-
-**In scope:**
- Per-game SQLite container (`game.db`) holding action history, state snapshots, and shardok results.
- A new `SqliteHistory` class implementing the existing `FullGameHistory` trait, drop-in replacement for `PersistedHistory` at the call site.
-
-**Out of scope (deferred or rejected):**
- **Migration of existing saves.** Pre-alpha there are only two users and a handful of test games. Cheaper to nuke the existing save directories at cutover than to write and verify a migration importer. New games created post-cutover land in `game.db` directly; existing games are gone.
- The top-level `games.e0es` running-games registry. Stays as-is; its cache (PR #6706) already works.
- Idle-game eviction. Enabled by this work but lands in a follow-up phase after `SqliteHistory` is the authoritative path.
- A shared cross-game analytics DB. Per-game design supports ad-hoc cross-game queries via iterate-and-aggregate; build the centralized analytics DB only if/when those queries become hot.
-
-## File layout
-
-`game.db` lives in the existing per-game save directory, alongside `text_store.db`:
-
-```
-${EAGLE_SAVE_DIR}/${gameIdHex}/
-├── game.db          # NEW — action history, snapshots, shardok results
-└── text_store.db    # existing — client text (SqliteClientTextStore)
-```
-
-Existing `.e0a` / `.e0s` / `.e0i` / `directory.e0i` files do not coexist with `game.db`: at cutover we nuke save directories one time. There is no migration path and no legacy-file fallback.
-
-### Why `game.db` and `text_store.db` are separate files (not tables in one DB)
-
-It's tempting to combine them — one file per game, one connection per game, one cloud-upload story. The reason not to: **writer-lock contention.** SQLite serializes writers per-database file. `SqliteClientTextStore` writes frequently during LLM streaming (one `UPDATE` per token append on `texts.text`). `SqliteHistory` writes per-action during turn commits. Combining them means an in-flight LLM token append can block a turn commit, or vice versa. Separate DBs have separate writer locks and never contend. Subsystem ownership (each package owning its own schema, independent migration paths if the schemas evolve) is a bonus.
-
-`Persister` integration follows the `SqliteClientTextStore` pattern: `game.db` is uploaded to / downloaded from cloud storage as a single opaque blob, keyed by the filename. On first load, if `game.db` is missing locally, try `persister.retrieveAsStream("game.db")`. If both are missing, this is a new game and we create an empty DB.
-
-## Schema
-
-```sql
-- Action results, one row per action_result_index.
-CREATE TABLE action_results (
-    action_seq           INTEGER PRIMARY KEY,        -- 0-based, dense, never sparse
-    action_result_type   INTEGER NOT NULL,           -- denormalized for filtering
-    round_id             INTEGER NOT NULL,           -- denormalized for recentResultsForRound
-    date_year            INTEGER,                    -- denormalized for sinceDate; NULL pre-game-start
-    date_month           INTEGER,                    -- denormalized for sinceDate; NULL pre-game-start
-    payload              BLOB    NOT NULL            -- ActionResult proto bytes
-) WITHOUT ROWID;
-
-CREATE INDEX idx_action_results_round   ON action_results(round_id);
-CREATE INDEX idx_action_results_date    ON action_results(date_year, date_month);
-- action_seq is the PK so no index needed there.
-
-- State snapshots at chunk boundaries (and the starting state at seq 0).
-- boundary_action_seq is the action_seq AFTER which the snapshot reflects state.
-- The starting state lives at boundary_action_seq = 0 (state before any actions).
-CREATE TABLE state_snapshots (
-    boundary_action_seq  INTEGER PRIMARY KEY,
-    payload              BLOB    NOT NULL            -- GameState proto bytes
-) WITHOUT ROWID;
-
-- Shardok per-battle results.
-CREATE TABLE shardok_results (
-    shardok_game_id      TEXT    NOT NULL,
-    action_seq           INTEGER NOT NULL,           -- 0-based within this shardok game
-    payload              BLOB    NOT NULL,           -- ShardokActionResult proto bytes
-    PRIMARY KEY (shardok_game_id, action_seq)
-) WITHOUT ROWID;
-
-- Shardok per-game state (one row per shardok_game_id).
-CREATE TABLE shardok_state (
-    shardok_game_id      TEXT    PRIMARY KEY,
-    game_state           BLOB    NOT NULL,           -- ShardokGameState proto bytes
-    last_eagle_round_id  INTEGER NOT NULL
-) WITHOUT ROWID;
-
-- Shardok per-player results (filtered ActionResultView per faction).
-CREATE TABLE shardok_player_results (
-    shardok_game_id      TEXT    NOT NULL,
-    faction_id           INTEGER NOT NULL,
-    seq                  INTEGER NOT NULL,
-    payload              BLOB    NOT NULL,           -- ShardokActionResultView proto bytes
-    PRIMARY KEY (shardok_game_id, faction_id, seq)
-) WITHOUT ROWID;
-
-- Shardok per-player available commands (latest only, keyed by game + faction).
-CREATE TABLE shardok_player_commands (
-    shardok_game_id      TEXT    NOT NULL,
-    faction_id           INTEGER NOT NULL,
-    payload              BLOB,                       -- ShardokAvailableCommands proto bytes; NULL = no commands
-    PRIMARY KEY (shardok_game_id, faction_id)
-) WITHOUT ROWID;
-
-- Schema version and starting state.
-CREATE TABLE metadata (
-    key                  TEXT    PRIMARY KEY,
-    value                BLOB    NOT NULL
-);
-- Seeded with: schema_version=1
-```
-
-### Column rationale
-
- **`action_seq`** as `INTEGER PRIMARY KEY` (the SQLite rowid alias) — densest possible storage and no separate index; range scans for `since(start)` are O(log n + result count).
- **`WITHOUT ROWID`** on tables with synthetic keys to skip the implicit rowid column.
- **Denormalized `action_result_type`, `round_id`, `date_year`, `date_month`** — these are the predicates the existing read paths use (`recentResultsForRound`, `sinceDate`). Keeping them as columns avoids parsing the proto blob to filter.
- **No `created_at` timestamp** — wall-clock time isn't queried by the History trait, and game date (year/month) is what matters semantically.
- **`payload BLOB`** — the proto bytes are the source of truth for the structured action result. Polymorphic action types make a relational schema painful for limited gain; the denormalized columns above cover the queries we need.
- **Snapshots keyed by `boundary_action_seq`** — `stateAfter(N)` finds `MAX(boundary_action_seq) WHERE boundary_action_seq <= N`, returns that snapshot, replays forward `N - boundary_action_seq` actions. Snapshot at `0` is the game's starting state.
-
-### Snapshot strategy
-
-Match the existing `resultsPerSaveFile = 25` boundary: write a `state_snapshots` row every 25 actions. That gives the same replay-window cost as the current chunk-file design (`stateAfter(N)` replays at most 25 actions to reach an arbitrary point), and matches the cadence developers are already calibrated to.
-
-Snapshots are GameState proto bytes, identical in shape to today's chunk-file `startingState`. The migration importer derives them directly from the chunk files. New games write a snapshot after every 25th `withNewResults` action.
-
-## Connection lifecycle
-
-Mirror `SqliteClientTextStore`:
-
- One `Connection` per loaded game, opened when `GamesManager` loads the game, closed when the game is evicted (future eviction work) or the server shuts down.
- `Class.forName("org.sqlite.JDBC")` + `DriverManager.getConnection("jdbc:sqlite:${path}")` on open.
- `PRAGMA journal_mode = WAL` on every connection open. WAL gives us crash-safe writes and lets readers proceed concurrently with the single writer — important because gRPC stream readers (humanPlayerClientConnectionState) query history mid-turn.
- `PRAGMA synchronous = NORMAL` (the WAL-recommended setting; durability is preserved through WAL checkpoint).
- `PRAGMA foreign_keys = ON` (defensive; we have no FKs today but cheap to enable).
- Auto-commit on by default. Transactions explicitly opened for `withNewResults` (batch of action inserts + optional snapshot) and `truncateTo` (deletes across all tables).
-
-### Threading
-
-The current `PersistedHistory` is an immutable case class; `withNewResults` returns a new instance. `SqliteHistory` cannot be pure-immutable (the DB is mutable state) but should present the same interface: methods that "change" the history return `this` after a successful write. The underlying `Connection` is shared.
-
-JDBC `Connection` is not thread-safe in general; SQLite's JDBC driver serializes operations per-connection. Existing call sites already serialize writes through the `EngineApplier` flow, so single-threaded write access is preserved. Reads from gRPC stream readers can use the same connection — SQLite serializes them transparently, and WAL prevents read-write blocking.
-
-## Read paths
-
-How each `FullGameHistory` method maps to SQL:
-
-| Method | Query |
-|---|---|
-| `count` | `SELECT COALESCE(MAX(action_seq), -1) + 1 FROM action_results` (cached as a counter after the first read) |
-| `last` | `SELECT payload FROM action_results ORDER BY action_seq DESC LIMIT 1` + state from `stateAfter(count)` |
-| `all` | `SELECT payload FROM action_results ORDER BY action_seq` — used by `GameAdminServiceImpl.getActionDetail` to fetch one action by index. See follow-up note below. |
-| `since(start)` | `SELECT payload FROM action_results WHERE action_seq >= ? ORDER BY action_seq` |
-| `sinceDate(date)` | `SELECT payload FROM action_results WHERE (date_year, date_month) >= (?, ?) ORDER BY action_seq` |
-| `recentResultsForRound(round, pred)` | `SELECT payload FROM action_results WHERE round_id = ? AND action_seq > ? ORDER BY action_seq` where `?` is the cutoff matching current `recentHistory` semantics (last N actions, or all actions for the current round) |
-| `stateAfter(N)` | Find latest snapshot ≤ N, replay forward via `replayApplier` (same logic as `replayScalaOnlyToState`) |
-
-For methods that return `Vector[ActionResultWithResultingState]` (with resulting state per row): we **do not** materialize per-row gameStates. Instead, fold the actions through `replayApplier` starting from the latest snapshot ≤ start, producing the states on the fly. This matches what `formAwrs` does today, with one critical difference: no `PersistedActionResult` wrapper is allocated, and no proto-conversion is performed per step. This is the same shape as `replayScalaOnlyToState`, generalized to produce intermediate states.
-
-### `all()` follow-up
-
-`GameAdminServiceImpl.getActionDetail` is the one production caller. It fetches a single action by index from the full vector. Either of these is cheaper than materializing the whole history:
-
- Replace the call site with `history.since(index).headOption`.
- Add a new `actionAt(index): Option[ActionResultWithResultingState]` method and drop `all()` from the trait entirely.
-
-`SqliteHistory.all` will work — it's just `SELECT * ORDER BY action_seq` — but it's expensive (materializes the full history into memory), so we should switch the admin caller in a small follow-up PR. Not blocking the SQLite work.
-
-### `recentResultsForRound` semantics
-
-Today this filters `recentHistory` (the in-memory tail) by `roundId`. In SQLite there's no in-memory/persisted split — all results live in the DB. The semantics shift slightly: return all results matching `round_id` after a configurable cutoff. The cutoff should match today's behavior (results since the start of the current round, or some bounded recent window). Default to "all results with the given `round_id`," which is correct as long as `round_id` uniquely identifies a round across the game's history (it does, per the current `RoundId` model).
-
-## Write paths
-
-### `withNewResults(newResults)`
-
-In a transaction:
-1. `INSERT INTO action_results (action_seq, action_result_type, round_id, date_year, date_month, payload) VALUES ...` — one row per new result. Use `addBatch()` for multiple.
-2. For each new result whose `action_seq % 25 == 0` (snapshot boundary), `INSERT INTO state_snapshots(boundary_action_seq, payload) VALUES (?, ?)` with the GameState proto bytes.
-3. Commit.
-
-The denormalized columns (`action_result_type`, `round_id`, `date_year`, `date_month`) are extracted from the action's resulting state at insert time. They are immutable once written; if the schema interpretation changes, a migration is required.
-
-The "individual result for crash recovery" pattern (`.e0i` files) is replaced by: the action result is durably written when the transaction commits. WAL gives us atomicity per transaction. No separate crash-recovery file is needed.
-
-### `saveNow`
-
-Becomes a no-op in normal operation — writes are already durable per `withNewResults` commit. We keep the method on the trait for API compatibility but the implementation just returns `this`. (We could call `PRAGMA wal_checkpoint(TRUNCATE)` here to roll the WAL into the main DB file, useful before cloud upload; defer this until we measure WAL growth.)
-
-### `truncateTo(targetActionCount)`
-
-In a transaction:
-1. `DELETE FROM action_results WHERE action_seq >= ?`
-2. `DELETE FROM state_snapshots WHERE boundary_action_seq > ?`
-3. Shardok cleanup: `DELETE FROM shardok_results / shardok_state / shardok_player_results / shardok_player_commands WHERE shardok_game_id NOT IN (...)` (the set of battles still outstanding at the truncate point; mirrors `deleteOrphanedShardokFiles`).
-4. Commit.
-
-`truncateTo(0)` resets the game; everything after the starting-state snapshot is deleted.
-
-## Shardok results
-
-The shardok subsystem currently lives in `.e0s` files (one per battle, full per-battle state). It's loaded selectively at game-load time, only for outstanding battles (see `PersistedHistory.apply` line 248-256).
-
-Moving to SQLite, the four shardok tables above capture:
- `shardok_results` — the per-battle result stream
- `shardok_state` — current per-battle state (one row per battle)
- `shardok_player_results` — per-faction filtered views
- `shardok_player_commands` — current available commands per faction
-
-`withNewShardokResults` writes to all four in a transaction. `shardokCount`, `shardokGameState`, etc., become single-row indexed lookups.
-
-The selective-load optimization disappears with SQLite: we don't proactively read anything; queries hit the DB on demand. The "load only outstanding battles" logic is replaced by "query by `shardok_game_id` when needed."
-
-## Crash recovery
-
-WAL replaces the `.e0i` individual-result-file mechanism. On startup:
- WAL is automatically replayed by SQLite if the previous shutdown was unclean. No application code needed.
- A successful commit means durable; an interrupted commit means rolled back. No half-written state visible.
-
-The current "orphaned individual results on load" path (`loadIndividualResults` in `PersistedHistory.apply`) is gone.
-
-## Cutover
-
-No migration path. At cutover:
-
-1. Stop the server.
-2. Nuke the contents of `${EAGLE_SAVE_DIR}` (and `${EAGLE_ARCHIVE_DIR}` if there's anything there).
-3. Deploy.
-4. New games created post-deploy land in `game.db` directly.
-
-Pre-alpha there are only two users and a handful of test games; a one-time nuke is cheaper than a verified migration importer. Trade-off accepted by the user explicitly.
-
-### Testing strategy (no migration)
-
-Without migration, we don't need equivalence-vs-`PersistedHistory` testing. The correctness gate becomes:
-
-1. **Adapt `PersistedHistoryTest`** to run against `SqliteHistory` instead. The 890-line existing test suite covers the trait surface exhaustively. Same assertions, new implementation under test.
-2. **Property-style end-to-end test**: create a synthetic game, run N batches of `withNewResults`, verify that `since(start)`, `stateAfter(N)`, `sinceDate(date)`, `recentResultsForRound`, `truncateTo`, etc., all produce expected values. Run with a deterministic random seed.
-3. **Manual smoke test in dev**: start a fresh game, play several turns, restart the server, verify saves persist and replay correctly.
-
-## Cloud-storage integration
-
-`Persister.save(key, bytes)` and `persister.retrieveAsStream(key)` already handle local-vs-S3 transparently. `SqliteClientTextStore` integrates by treating `text_store.db` as a single binary blob: upload after writes, download before reads.
-
-`SqliteHistory` will do the same with `game.db`:
- **On load:** if local `game.db` missing, try to download via `persister.retrieveAsStream("game.db")`. Fall back to migration from `.e0a` if both are missing.
- **On save:** after a meaningful change (turn boundary? configurable cadence?), upload the current `game.db` file to S3. Need to decide the upload trigger — too frequent and we burn S3 PUTs; too rare and crash recovery is lossy.
-
-**Open question:** the right upload cadence. Today, chunk saves trigger cloud upload at chunk boundaries (every 25 actions). The simplest match: keep that cadence, upload `game.db` once per chunk boundary. We could be smarter (only upload if WAL has been checkpointed; only upload deltas) but those are optimizations.
-
-## What the trait-level cutover looks like
-
-`GamesManager` currently has roughly:
-
-```scala
-val history: FullGameHistory = PersistedHistory(gameId, persister).getOrElse(...)
-```
-
-Becomes:
-
-```scala
-val history: FullGameHistory = SqliteHistory.loaded(gameId, persister)
-```
-
-Every other consumer of `FullGameHistory` is unchanged.
-
-## Decisions locked
-
-1. **Per-game `game.db`** (not shared across games).
-2. **`game.db` and `text_store.db` stay separate files** (not combined into one DB). Writer-lock contention is the deciding factor.
-3. **No migration** of existing saves. Nuke save directories at cutover. Pre-alpha trade-off accepted by the user.
-4. **No feature flag.** Cutover is unconditional. Failures are loud.
-5. **Cloud upload cadence**: match the existing 25-action chunk cadence. Upload `game.db` once per snapshot boundary.
-6. **`recentResultsForRound`**: return all results for the given round (no recent-window cap). The `round_id` predicate bounds the scan naturally.
-7. **WAL checkpointing**: rely on SQLite's auto-checkpoint. Add explicit `PRAGMA wal_checkpoint` only if WAL file growth becomes a problem.
-8. **Schema versioning**: seed `metadata` with `schema_version = 1`. Schema migration mechanism deferred until we need a v2.
-9. **`history.all`**: keep on the trait for now; admin call site moved to `since(i).headOption` (or a new `actionAt(i)` method) in a small follow-up PR. Not blocking.
-
-## Phase plan (revised)
-
-| Phase | Work | Estimate |
-|---|---|---|
-| 0 | This design doc | 2-3 days (in flight) |
-| 1 | `SqliteHistory` schema + write paths + tests adapted from `PersistedHistoryTest` | ~1 week |
-| 2 | Read paths + full test parity | ~1 week |
-| 3 | Shardok results integration | 3-5 days |
-| 4 | Cutover in `GamesManager` (nuke `${EAGLE_SAVE_DIR}` as part of deploy) | 2-3 days |
-| 5 | Idle-game eviction | 3-5 days |
-| 6 | Post-alpha cleanup (delete `PersistedHistory`, `PersistedActionResult`, `PartialGameUtils`, the chunk-file save code, `games.e0es` cache if no longer needed) | 1-2 days |
-
-**Total: ~3-4 weeks** (down from 5-6 — migration and equivalence-testing dropped).
-
-The follow-up to migrate `GameAdminServiceImpl.getActionDetail` off `history.all` is a small, independent PR that can land any time.
@@ -22,7 +22,7 @@ This avoids the "everything is pink" problem of switching the pipeline on main b
 | Shaders | 47 | 7 custom Eagle, 1 hex mesh, 20 Polytope, 13 TextMesh Pro, 6 other third-party |
 | Materials | 325 | Mix of Standard shader, custom, and third-party |
 | ShaderGraph files | 4 | All TextMesh Pro (URP + HDRP variants already exist) |
-| Scenes with baked lighting | 5 | Connection, Eagle, Shardok, Shared, Map Editor |
+| Scenes with baked lighting | 6 | Gameplay, Connection, Eagle, Shardok, Shared, Map Editor |
 | Post-processing | 1 profile | PPP_Orc.asset (Post Processing Stack v2, FXAA/TAA, AO) |
 | C# rendering scripts | 16 | Material/shader property manipulation only |
 | Rendering path | Forward | Matches URP default |
@@ -132,7 +132,7 @@ This is the critical path. Without these, the game is unplayable.
 - Fix any visual differences from lighting model changes

 **Lighting**
- Rebake lightmaps for all 5 scenes (Connection, Eagle, Shardok, Shared, Map Editor)
+- Rebake lightmaps for all 6 scenes (Gameplay, Connection, Eagle, Shardok, Shared, Map Editor)
 - Configure URP shadow cascade settings to match current quality levels
 - Verify HDR rendering

@@ -4,18 +4,11 @@ This document describes which quests the AI attempts to complete proactively to

 ## Overview

-The AI attempts to complete most quests via `FulfillQuestsCommandSelector`, which is invoked by `MidGameAIClient.chosenFulfillEasyQuestsCommand`.
-
-Important scope limits:
-
- The selector only considers quests from unaffiliated heroes in provinces owned by the acting faction.
- It further filters to provinces ruled by that faction's leader.
- It only emits a command when the corresponding command is currently available.
- Handler order is fixed. The first handler that produces a valid command wins.
+The AI attempts to complete quests via `FulfillQuestsCommandSelector`, which is invoked by `MidGameAIClient.chosenFulfillEasyQuestsCommand`. The AI only considers quests from unaffiliated heroes in provinces ruled by a faction leader.

 ## Quests the AI Actively Completes

-This section lists quests with direct `QuestCommandChooser` handlers in `FulfillQuestsCommandSelector`.
+The AI processes quest handlers in priority order. The first handler that produces a valid command wins.

 ### Diplomacy Quests

@@ -43,8 +36,7 @@ This section lists quests with direct `QuestCommandChooser` handlers in `Fulfill
 | `AlmsAcrossRealmQuest` | `AlmsAcrossRealmQuestCommandChooser` | Gives food from the province with the largest surplus |
 | `GiveToHeroesInProvinceQuest` | `GiveToHeroesInProvinceQuestCommandChooser` | Gives gold to the hero with the lowest loyalty in the specified province |
 | `GiveToHeroesAcrossRealmQuest` | `GiveToHeroesAcrossRealmQuestCommandChooser` | Gives gold from the province with the most gold available |
-| `SpendOnFeastsInProvinceQuest` | `SpendOnFeastsQuestCommandChooser` | Spend gold on feasts in the quest holder's province |
-| `SpendOnFeastsAcrossRealmQuest` | `SpendOnFeastsQuestCommandChooser` | Spend gold on the most expensive currently available feast |
+| `SpendOnFeastsQuest` | `SpendOnFeastsQuestCommandChooser` | Spend gold on feasts |
 | `SendSuppliesQuest` | `SendSuppliesQuestCommandChooser` | Send food to the target province |

 ### Prisoner Quests
@@ -80,6 +72,12 @@ This section lists quests with direct `QuestCommandChooser` handlers in `Fulfill
 | `BattalionDiversityQuest` | `BattalionDiversityQuestCommandChooser` | Hires one battalion of a missing type to reach 3+ distinct types. Only attempts if the province already has 2+ existing types, has sufficient development (`meetsRequirements`), and gold/food surplus. |
 | `UpgradeBattalionQuest` | `UpgradeBattalionQuestCommandChooser` | Walks a priority tree per turn: arm-completes (with optional travel and/or food sale setup) → train-completes → march in a pre-qualified neighbor battalion → organize/top-off of the target type → march in an unqualified neighbor battalion → develop Economy/Agriculture → train progress → develop Infrastructure → arm progress. Only toggles Travel when it directly enables an arm-completes finisher. |

+### Riot Quests
+
+| Quest | Handler | Notes |
+|-------|---------|-------|
+| `SuppressRiotByForceQuest` | `CommandChoiceHelpers.handleRiotSelectedCommand` | When this quest exists and the faction has battalions, the AI prefers CrackDown over Give when handling riots. Not a quest command chooser — modifies existing riot handling priority. |
+
 ### Reconnaissance Quests

 | Quest | Handler | Notes |
@@ -107,14 +105,6 @@ This section lists quests with direct `QuestCommandChooser` handlers in `Fulfill
 | `SwearBrotherhoodWithHeroQuest` | `SwearBrotherhoodQuestCommandChooser` | Swear brotherhood with a specific hero |
 | `DismissSpecificVassalQuest` | `DismissSpecificVassalCommandChooser` | Only if province has more than 2 heroes AND the unaffiliated hero's power >= target hero's power * `RequiredPowerMultiplierForDismiss` |

-## Quests the AI Handles Indirectly
-
-These are not in `FulfillQuestsCommandSelector`, but other AI command-selection code can still bias toward completing them.
-
-| Quest | Handler | Notes |
-|-------|---------|-------|
-| `SuppressRiotByForceQuest` | `CommandChoiceHelpers.handleRiotSelectedCommand` | When this quest exists and the faction has battalions, the AI prefers CrackDown over Give when handling riots. |
-
 ## Quests the AI Does Not Yet Attempt to Complete

 The following quests have no handler and must be completed naturally through gameplay. They are listed in rough priority order for future implementation.
@@ -163,7 +153,7 @@ The AI prioritizes quests in the order they appear in `FulfillQuestsCommandSelec
 13. ReturnPrisoner
 14. ReleaseAllPrisoners
 15. ApprehendOutlaw
-16. SpendOnFeastsInProvince / SpendOnFeastsAcrossRealm
+16. SpendOnFeasts
 17. RestProvince
 18. DevelopProvinces
 19. MobilizeProvinces
@@ -87,7 +87,7 @@ Create these prefabs in `Assets/Eagle/Effects/`:

 ### Scene Setup

-Wire up in `Assets/Scenes/Eagle.unity`:
+Wire up in `Gameplay.unity`:
 1. Add controller components to the Map GameObject
 2. Assign `mapContainer`, `centroidsJson` (shared with other controllers)
 3. Assign the effect prefab for each controller
@@ -2,7 +2,7 @@ module github.com/nolen777/eagle0

 go 1.23.0

-toolchain go1.25.11
+toolchain go1.23.3

 require (
 	github.com/aws/aws-sdk-go-v2 v1.32.8
@@ -12,9 +12,9 @@ require (
 	github.com/golang-jwt/jwt/v5 v5.2.1
 	github.com/google/uuid v1.6.0
 	github.com/webview/webview_go v0.0.0-20240831120633-6173450d4dd6
-	golang.org/x/sys v0.33.0
-	google.golang.org/grpc v1.71.0
-	google.golang.org/protobuf v1.36.10
+	golang.org/x/sys v0.30.0
+	google.golang.org/grpc v1.68.0
+	google.golang.org/protobuf v1.36.3
 )

 require (
@@ -32,7 +32,5 @@ require (
 	github.com/aws/aws-sdk-go-v2/service/ssooidc v1.28.8 // indirect
 	github.com/aws/aws-sdk-go-v2/service/sts v1.33.6 // indirect
 	github.com/aws/smithy-go v1.22.1 // indirect
-	golang.org/x/net v0.34.0 // indirect
 	golang.org/x/text v0.25.0 // indirect
-	google.golang.org/genproto/googleapis/rpc v0.0.0-20250115164207-1a7da9e5054f // indirect
 )
@@ -43,24 +43,14 @@ github.com/google/uuid v1.6.0 h1:NIvaJDMOsjHA8n1jAhLSgzrAzy1Hgr+hNrb57e+94F0=
 github.com/google/uuid v1.6.0/go.mod h1:TIyPZe4MgqvfeYDBFedMoGGpEw/LqOeaOT+nhxU+yHo=
 github.com/webview/webview_go v0.0.0-20240831120633-6173450d4dd6 h1:VQpB2SpK88C6B5lPHTuSZKb2Qee1QWwiFlC5CKY4AW0=
 github.com/webview/webview_go v0.0.0-20240831120633-6173450d4dd6/go.mod h1:yE65LFCeWf4kyWD5re+h4XNvOHJEXOCOuJZ4v8l5sgk=
-golang.org/x/net v0.34.0 h1:Mb7Mrk043xzHgnRM88suvJFwzVrRfHEHJEl5/71CKw0=
-golang.org/x/net v0.34.0/go.mod h1:di0qlW3YNM5oh6GqDGQr92MyTozJPmybPK4Ev/Gm31k=
 golang.org/x/sys v0.30.0 h1:QjkSwP/36a20jFYWkSue1YwXzLmsV5Gfq7Eiy72C1uc=
 golang.org/x/sys v0.30.0/go.mod h1:/VUhepiaJMQUp4+oa/7Zr1D23ma6VTLIYjOOTFZPUcA=
-golang.org/x/sys v0.33.0 h1:q3i8TbbEz+JRD9ywIRlyRAQbM0qF7hu24q3teo2hbuw=
-golang.org/x/sys v0.33.0/go.mod h1:BJP2sWEmIv4KK5OTEluFJCKSidICx8ciO85XgH3Ak8k=
 golang.org/x/text v0.25.0 h1:qVyWApTSYLk/drJRO5mDlNYskwQznZmkpV2c8q9zls4=
 golang.org/x/text v0.25.0/go.mod h1:WEdwpYrmk1qmdHvhkSTNPm3app7v4rsT8F2UD6+VHIA=
 golang.org/x/xerrors v0.0.0-20191204190536-9bdfabe68543/go.mod h1:I/5z698sn9Ka8TeJc9MKroUUfqBBauWjQqLJ2OPfmY0=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20250115164207-1a7da9e5054f h1:OxYkA3wjPsZyBylwymxSHa7ViiW1Sml4ToBrncvFehI=
-google.golang.org/genproto/googleapis/rpc v0.0.0-20250115164207-1a7da9e5054f/go.mod h1:+2Yz8+CLJbIfL9z73EW45avw8Lmge3xVElCP9zEKi50=
 google.golang.org/grpc v1.68.0 h1:aHQeeJbo8zAkAa3pRzrVjZlbz6uSfeOXlJNQM0RAbz0=
 google.golang.org/grpc v1.68.0/go.mod h1:fmSPC5AsjSBCK54MyHRx48kpOti1/jRfOlwEWywNjWA=
-google.golang.org/grpc v1.71.0 h1:kF77BGdPTQ4/JZWMlb9VpJ5pa25aqvVqogsxNHHdeBg=
-google.golang.org/grpc v1.71.0/go.mod h1:H0GRtasmQOh9LkFoCPDu3ZrwUtD1YGE+b2vYBYd/8Ec=
 google.golang.org/protobuf v1.26.0-rc.1 h1:7QnIQpGRHE5RnLKnESfDoxm2dTapTZua5a0kS0A+VXQ=
 google.golang.org/protobuf v1.26.0-rc.1/go.mod h1:jlhhOSvTdKEhbULTjvd4ARK9grFBp09yW+WbY/TyQbw=
 google.golang.org/protobuf v1.36.3 h1:82DV7MYdb8anAVi3qge1wSnMDrnKK7ebr+I0hHRN1BU=
 google.golang.org/protobuf v1.36.3/go.mod h1:9fA7Ob0pmnwhb644+1+CVWFRbNajQ6iRojtC/QF5bRE=
-google.golang.org/protobuf v1.36.10 h1:AYd7cD/uASjIL6Q9LiTjz8JLcrh/88q5UObnmY3aOOE=
-google.golang.org/protobuf v1.36.10/go.mod h1:HTf+CrKn2C3g5S8VImy6tdcUvCska2kB7j23XfzDpco=
@@ -4,10 +4,8 @@

 set -e

-export PATH="/opt/homebrew/bin:/usr/local/bin:$PATH"
-
 # Run gazelle
-bazel run //:gazelle
+bazel run //:gazelle 2>/dev/null

 # Check if any BUILD files were modified
 if ! git diff --quiet -- '*.bazel' '**/BUILD' 'WORKSPACE*'; then
@@ -1,22 +0,0 @@
-#!/bin/bash
-# Pre-commit hook wrapper for scalafmt.
-# Runs the repository-pinned scalafmt CLI through Bazel.
-
-set -e
-
-export PATH="/opt/homebrew/bin:/usr/local/bin:$PATH"
-
-if [ "$#" -eq 0 ]; then
-    exit 0
-fi
-
-repo_root="$(git rev-parse --show-toplevel)"
-files=()
-for file in "$@"; do
-    case "$file" in
-        /*) files+=("$file") ;;
-        *) files+=("$repo_root/$file") ;;
-    esac
-done
-
-bazel run //tools:scalafmt -- --config "$repo_root/.scalafmt.conf" -i "${files[@]}"
@@ -1,12 +1,11 @@
 #!/usr/bin/env bash
-# Keeps Bazel macOS builds in sync with the installed Xcode version.
+# Keeps Bazel mactools builds in sync with the installed Xcode version.
 #
-# Needed on macOS runners and developer machines because Bazel's generated
-# local_config_xcode/local_config_apple_cc repositories bake in the discovered
-# Xcode version. If Xcode is upgraded in place, the DEVELOPER_DIR path stays the
-# same, so Bazel can otherwise keep using stale Xcode metadata.
+# Only needed on runners that do mactools builds (Mac/Sparkle). Non-Mac builds
+# don't run this script; their apple_cc_autoconf stays cached because none of
+# its environ vars change, so they're unaffected by Xcode version changes.
 #
-# 1. Writes .bazelrc.xcode with macOS-scoped flags:
+# 1. Writes .bazelrc.xcode with mactools-scoped flags:
 #    - action_env for remote cache key invalidation
 #    - DEVELOPER_DIR pointing to the active Xcode
 #
@@ -18,24 +17,17 @@

 set -euo pipefail

-DEVELOPER_DIR=""
-if [ -d "/Applications/Xcode.app/Contents/Developer" ]; then
-    DEVELOPER_DIR="/Applications/Xcode.app/Contents/Developer"
-else
-    DEVELOPER_DIR=$(xcode-select -p 2>/dev/null || true)
-fi
-
-XCODE_BUILD_VERSION=$(DEVELOPER_DIR="$DEVELOPER_DIR" xcodebuild -version 2>/dev/null | awk '/Build version/ {print $3}' || true)
+XCODE_BUILD_VERSION=$(xcodebuild -version 2>/dev/null | awk '/Build version/ {print $3}')

 if [ -z "$XCODE_BUILD_VERSION" ]; then
-    echo "ERROR: Could not detect full Xcode build version."
-    echo "mactools/Sparkle builds require full Xcode, not just Command Line Tools."
-    echo "Install Xcode at /Applications/Xcode.app or select it with xcode-select."
-    exit 1
+    echo "Warning: Could not detect Xcode build version, skipping"
+    exit 0
 fi

+DEVELOPER_DIR=$(xcode-select -p 2>/dev/null)
+
 BAZELRC_XCODE=".bazelrc.xcode"
-EXPECTED_LINE="common:macos --action_env=XCODE_BUILD_VERSION=${XCODE_BUILD_VERSION}"
+EXPECTED_LINE="common:mactools --action_env=XCODE_BUILD_VERSION=${XCODE_BUILD_VERSION}"

 # Check if the file already has the right version (first line is the version marker)
 if [ -f "$BAZELRC_XCODE" ]; then
@@ -57,11 +49,6 @@ bazel clean --expunge 2>/dev/null || true

 cat > "$BAZELRC_XCODE" << EOF
 ${EXPECTED_LINE}
-common:macos --repo_env=DEVELOPER_DIR=${DEVELOPER_DIR}
+common:mactools --repo_env=DEVELOPER_DIR=${DEVELOPER_DIR}
 EOF
-
-if [ -n "${GITHUB_ENV:-}" ]; then
-    echo "DEVELOPER_DIR=${DEVELOPER_DIR}" >> "$GITHUB_ENV"
-fi
-
-echo "Updated ${BAZELRC_XCODE} — next macOS Bazel build will rebuild with Xcode ${XCODE_BUILD_VERSION}"
+echo "Updated ${BAZELRC_XCODE} — next mactools build will rebuild with Xcode ${XCODE_BUILD_VERSION}"
@@ -66,14 +66,6 @@ cc_library(
    ],
 )

-cc_library(
-    name = "shardok_latency_trace",
-    srcs = ["ShardokLatencyTrace.cpp"],
-    hdrs = ["ShardokLatencyTrace.hpp"],
-    copts = COPTS,
-    visibility = ["//visibility:public"],
-)
-
 cc_library(
    name = "repeated_proto_utils",
    srcs = [],
@@ -88,11 +88,6 @@ StdLibraryGenerator::StdLibraryGenerator() : RandomGenerator() {
    engine.seed(static_cast<std::mt19937_64::result_type>(std::time(nullptr)));
 }

-StdLibraryGenerator::StdLibraryGenerator(const std::mt19937_64::result_type seed)
-    : RandomGenerator() {
-    engine.seed(seed);
-}
-
 auto StdLibraryGenerator::IntBetween(const int min, const int max) -> int {
    std::uniform_int_distribution<int> unifInt(min, max - 1);
    return unifInt(engine);
@@ -51,7 +51,6 @@ private:

 public:
    StdLibraryGenerator();
-    explicit StdLibraryGenerator(std::mt19937_64::result_type seed);

    auto IntBetween(int min, int max) -> int override;  // inclusive, exclusive (min<=return<max)
 };
@@ -1,101 +0,0 @@
-#include "ShardokLatencyTrace.hpp"
-
-#include <cstdlib>
-#include <iostream>
-#include <mutex>
-#include <sstream>
-
-namespace net::eagle0::common {
-namespace {
-
-auto ReadEnvEnabled() -> bool {
-    const char* value = std::getenv("EAGLE0_SHARDOK_LATENCY_TRACE");
-    if (value == nullptr) { return false; }
-    const std::string text(value);
-    return text == "1" || text == "true" || text == "TRUE" || text == "yes" || text == "YES";
-}
-
-auto TraceEnabled() -> bool {
-    static const bool enabled = ReadEnvEnabled();
-    return enabled;
-}
-
-auto ReadGameFilter() -> std::string {
-    const char* value = std::getenv("EAGLE0_SHARDOK_LATENCY_GAME_ID");
-    return value == nullptr ? "" : std::string(value);
-}
-
-auto GameFilter() -> const std::string& {
-    static const std::string filter = ReadGameFilter();
-    return filter;
-}
-
-auto JsonEscape(std::string_view value) -> std::string {
-    std::ostringstream out;
-    for (const char c : value) {
-        switch (c) {
-            case '"': out << "\\\""; break;
-            case '\\': out << "\\\\"; break;
-            case '\n': out << "\\n"; break;
-            case '\r': out << "\\r"; break;
-            case '\t': out << "\\t"; break;
-            default: out << c; break;
-        }
-    }
-    return out.str();
-}
-
-auto LogMutex() -> std::mutex& {
-    static auto mutex = new std::mutex();
-    return *mutex;
-}
-
-}  // namespace
-
-auto ShardokLatencyTraceEnabledForGame(std::string_view gameId) -> bool {
-    if (!TraceEnabled()) { return false; }
-    const std::string& filter = GameFilter();
-    return filter.empty() || filter == gameId;
-}
-
-void EmitShardokLatencyTrace(
-        std::string_view event,
-        std::string_view gameId,
-        const ShardokLatencyFields& fields,
-        std::chrono::steady_clock::duration duration) {
-    if (!ShardokLatencyTraceEnabledForGame(gameId)) { return; }
-
-    const auto durationMs =
-            std::chrono::duration_cast<std::chrono::microseconds>(duration).count() / 1000.0;
-
-    std::ostringstream out;
-    out << "{\"event\":\"shardok_latency\","
-        << "\"component\":\"shardok\","
-        << "\"span\":\"" << JsonEscape(event) << "\","
-        << "\"game_id\":\"" << JsonEscape(gameId) << "\","
-        << "\"duration_ms\":" << durationMs;
-    for (const auto& [key, value] : fields) {
-        out << ",\"" << JsonEscape(key) << "\":\"" << JsonEscape(value) << "\"";
-    }
-    out << "}";
-
-    std::scoped_lock lock(LogMutex());
-    std::cerr << out.str() << std::endl;
-}
-
-ScopedShardokLatencyTrace::ScopedShardokLatencyTrace(
-        std::string event,
-        std::string gameId,
-        ShardokLatencyFields fields)
-    : event_(std::move(event)),
-      gameId_(std::move(gameId)),
-      fields_(std::move(fields)) {
-    if (ShardokLatencyTraceEnabledForGame(gameId_)) { start_ = std::chrono::steady_clock::now(); }
-}
-
-ScopedShardokLatencyTrace::~ScopedShardokLatencyTrace() {
-    if (!start_.has_value()) { return; }
-    EmitShardokLatencyTrace(event_, gameId_, fields_, std::chrono::steady_clock::now() - *start_);
-}
-
-}  // namespace net::eagle0::common
@@ -1,42 +0,0 @@
-#pragma once
-
-#include <chrono>
-#include <optional>
-#include <string>
-#include <string_view>
-#include <utility>
-#include <vector>
-
-namespace net::eagle0::common {
-
-using ShardokLatencyFields = std::vector<std::pair<std::string, std::string>>;
-
-auto ShardokLatencyTraceEnabledForGame(std::string_view gameId) -> bool;
-
-void EmitShardokLatencyTrace(
-        std::string_view event,
-        std::string_view gameId,
-        const ShardokLatencyFields& fields,
-        std::chrono::steady_clock::duration duration);
-
-class ScopedShardokLatencyTrace {
-public:
-    ScopedShardokLatencyTrace(
-            std::string event,
-            std::string gameId,
-            ShardokLatencyFields fields = {});
-    ~ScopedShardokLatencyTrace();
-
-    ScopedShardokLatencyTrace(const ScopedShardokLatencyTrace&) = delete;
-    auto operator=(const ScopedShardokLatencyTrace&) -> ScopedShardokLatencyTrace& = delete;
-    ScopedShardokLatencyTrace(ScopedShardokLatencyTrace&&) = delete;
-    auto operator=(ScopedShardokLatencyTrace&&) -> ScopedShardokLatencyTrace& = delete;
-
-private:
-    std::string event_;
-    std::string gameId_;
-    ShardokLatencyFields fields_;
-    std::optional<std::chrono::steady_clock::time_point> start_;
-};
-
-}  // namespace net::eagle0::common
@@ -21,14 +21,7 @@ namespace shardok::mcts {

 AbstractMCTSAI::AbstractMCTSAI(MCTSPlayerId playerId, MCTSConfig config)
    : playerId_(playerId),
-      config_(config),
-      randomEngine_(config.randomSeed == 0 ? std::random_device{}() : config.randomSeed) {}
-
-void AbstractMCTSAI::SetConfig(const MCTSConfig& newConfig) {
-    config_ = newConfig;
-    std::lock_guard lock(randomEngineMutex_);
-    randomEngine_.seed(newConfig.randomSeed == 0 ? std::random_device{}() : newConfig.randomSeed);
-}
+      config_(config) {}

 auto AbstractMCTSAI::Search(
        const MCTSGameEngine& engine,
@@ -357,19 +350,22 @@ auto AbstractMCTSAI::MCTSExpansion(MCTSNode* node, const MCTSGameEngine& engine)
        // fair UCB comparison with non-chance actions like END_TURN
        {
            double expectedImmediate = 0.0;
-            for (size_t i = 0; i < node->outcomeProbabilities.size(); i++) {
+            double totalProbability = 0.0;
+            for (size_t i = 0; i < node->children.size(); i++) {
                const double prob = node->outcomeProbabilities[i];
-                const double childImmediate = i < node->children.size()
-                                                      ? node->children[i]->immediateScore
-                                                      : node->unvisitedOutcomeScore;
+                const double childImmediate = node->children[i]->immediateScore;
                expectedImmediate += prob * childImmediate;
+                totalProbability += prob;
+            }
+            // Normalize by total probability of expanded outcomes
+            if (totalProbability > 0.0) {
+                node->immediateScore = expectedImmediate / totalProbability;
+                // CRITICAL: Always update lookaheadScore to the expected value.
+                // Without this, chance nodes keep their initial lookaheadScore from the parent
+                // state (before the action), while regular actions use the child state (after).
+                // This gives chance nodes an unfair initial UCB advantage.
+                node->lookaheadScore = node->immediateScore;
            }
-            node->immediateScore = expectedImmediate;
-            // CRITICAL: Always update lookaheadScore to the expected value.
-            // Without this, chance nodes keep their initial lookaheadScore from the parent
-            // state (before the action), while regular actions use the child state (after).
-            // This gives chance nodes an unfair initial UCB advantage.
-            node->lookaheadScore = node->immediateScore;
        }

        return node->children.back().get();
@@ -400,14 +396,8 @@ auto AbstractMCTSAI::MCTSExpansion(MCTSNode* node, const MCTSGameEngine& engine)

    const auto& action = nodeActions[actionIndex];

-    double totalActionWeight = 0.0;
-    for (const double weight : actionWeights) { totalActionWeight += std::max(0.0, weight); }
-
-    const double rawActionWeight =
-            actionIndex < actionWeights.size() ? std::max(0.0, actionWeights[actionIndex]) : 1.0;
-    const double actionWeight = totalActionWeight > 0.0
-                                        ? rawActionWeight / totalActionWeight
-                                        : 1.0 / static_cast<double>(nodeActions.size());
+    const double actionWeight =
+            actionIndex < actionWeights.size() ? actionWeights[actionIndex] : 1.0;

    // Check if this action requires a chance node
    if (action->requiresChanceNode()) {
@@ -427,16 +417,15 @@ auto AbstractMCTSAI::MCTSExpansion(MCTSNode* node, const MCTSGameEngine& engine)
        // Get outcome information from engine
        const auto outcomeInfo = engine.getBinaryOutcomeInfo(*node->gameState, *action);

-        // Set up outcome metadata.
+        // Set up outcome metadata (2 outcomes for binary actions)
        chanceNode->outcomeProbabilities = outcomeInfo.getProbabilities();
        chanceNode->outcomeRolls = outcomeInfo.getRepresentativeRolls();
-        chanceNode->totalActions = chanceNode->outcomeProbabilities.size();
+        chanceNode->totalActions = 2;  // Binary: success and failure

        // Chance node immediate score will be computed as expected value during backpropagation
        // For now, initialize to parent's score as a reasonable default
        chanceNode->immediateScore = engine.evaluateState(*node->gameState, playerId_);
        chanceNode->lookaheadScore = chanceNode->immediateScore;
-        chanceNode->unvisitedOutcomeScore = chanceNode->immediateScore;

        // Set parent and add to children
        chanceNode->parent = node;
@@ -613,20 +602,29 @@ auto AbstractMCTSAI::MCTSBackpropagation(
            // Chance nodes: compute expected value (weighted average of outcomes)
            // lookaheadScore = sum(probability[i] * childValue[i])
            double expectedValue = 0.0;
-            bool hasVisitedChild = false;
+            double totalProbability = 0.0;
+            int visitedChildCount = 0;
+
+            for (size_t i = 0; i < node->children.size(); i++) {
+                const auto& child = node->children[i];
+                if (child->visitCount == 0) continue;  // Unvisited outcomes don't contribute

-            for (size_t i = 0; i < node->outcomeProbabilities.size(); i++) {
                const double probability = node->outcomeProbabilities[i];
-                double childValue = node->unvisitedOutcomeScore;
-                if (i < node->children.size() && node->children[i]->visitCount > 0) {
-                    childValue = node->children[i]->lookaheadScore;
-                    hasVisitedChild = true;
-                }
+                const double childValue = child->lookaheadScore;
                expectedValue += probability * childValue;
+                totalProbability += probability;
+                visitedChildCount++;
            }

            // Use expected value if we have visited outcomes, else use average
-            if (hasVisitedChild) {
+            if (visitedChildCount > 0) {
+                // CRITICAL: Normalize by total probability to get correct expected value
+                // when not all outcomes have been visited yet
+                if (totalProbability > 0.0 && totalProbability < 1.0) {
+                    // Normalize to account for unvisited outcomes
+                    // This gives the correct expected value among visited outcomes
+                    expectedValue /= totalProbability;
+                }
                node->lookaheadScore = expectedValue;
            } else {
                // No outcomes visited yet, fall back to average
@@ -707,21 +705,23 @@ auto AbstractMCTSAI::SelectSimulationAction(
                "MCTS tree building or game state");
    }

+    thread_local std::mt19937 gen(std::random_device{}());
+
    switch (config_.simulationPolicy) {
        case MCTSSimulationPolicy::RANDOM: {
            std::uniform_int_distribution<size_t> dis(0, actions.size() - 1);
-            return SampleDistribution(dis);
+            return dis(gen);
        }

        case MCTSSimulationPolicy::FILTERED_RANDOM: {
            if (const auto filteredIndices = engine.filterActions(actions, state);
                !filteredIndices.empty()) {
                std::uniform_int_distribution<size_t> dis(0, filteredIndices.size() - 1);
-                return filteredIndices[SampleDistribution(dis)];
+                return filteredIndices[dis(gen)];
            }
            // Fall back to random
            std::uniform_int_distribution<size_t> dis(0, actions.size() - 1);
-            return SampleDistribution(dis);
+            return dis(gen);
        }

        case MCTSSimulationPolicy::BEST_IMMEDIATE: {
@@ -790,7 +790,7 @@ auto AbstractMCTSAI::SelectSimulationAction(

            // Select based on weights
            std::discrete_distribution<> dis(weights.begin(), weights.end());
-            return scores[SampleDistribution(dis)].first;
+            return scores[dis(gen)].first;
        }

        case MCTSSimulationPolicy::WEIGHTED_HEURISTIC: {
@@ -820,13 +820,13 @@ auto AbstractMCTSAI::SelectSimulationAction(

            // Select based on heuristic weights using discrete_distribution
            std::discrete_distribution<> dis(validWeights.begin(), validWeights.end());
-            return validIndices[SampleDistribution(dis)];
+            return validIndices[dis(gen)];
        }
    }

    // Default to random
    std::uniform_int_distribution<size_t> dis(0, actions.size() - 1);
-    return SampleDistribution(dis);
+    return dis(gen);
 }

 auto AbstractMCTSAI::FindNodeAtDepthWithHash(
@@ -7,8 +7,6 @@

 #include <chrono>
 #include <memory>
-#include <mutex>
-#include <random>
 #include <unordered_map>
 #include <vector>

@@ -43,7 +41,7 @@ public:

    // Configuration
    [[nodiscard]] auto GetConfig() const -> const MCTSConfig& { return config_; }
-    void SetConfig(const MCTSConfig& newConfig);
+    void SetConfig(const MCTSConfig& newConfig) { config_ = newConfig; }

    [[nodiscard]] auto FindNodeAtDepthWithHash(
            const MCTSNode* root,
@@ -53,8 +51,6 @@ public:
 private:
    MCTSPlayerId playerId_;
    MCTSConfig config_;
-    mutable std::mt19937 randomEngine_;
-    mutable std::mutex randomEngineMutex_;

    // Transposition table: maps state hash -> minimum depth at which state was reached
    // Used to detect and penalize longer paths to the same game state
@@ -89,13 +85,6 @@ private:
            const std::vector<std::unique_ptr<MCTSAction>>& actions,
            bool isMaximizing) const -> size_t;

-    template<typename Distribution>
-    [[nodiscard]] auto SampleDistribution(Distribution& distribution) const ->
-            typename Distribution::result_type {
-        std::lock_guard lock(randomEngineMutex_);
-        return distribution(randomEngine_);
-    }
-
    // Logging
    static auto LogSearchResults(
            const MCTSNode* rootNode,
@@ -114,4 +103,4 @@ private:
 }  // namespace mcts
 }  // namespace shardok

-#endif  // EAGLE0_ABSTRACT_MCTSAI_HPP
+#endif  // EAGLE0_ABSTRACT_MCTSAI_HPP
@@ -6,7 +6,6 @@ cc_library(
    copts = COPTS,
    visibility = [
        "//src/main/cpp/net/eagle0/shardok/ai/mcts:__subpackages__",
-        "//src/main/cpp/net/eagle0/shardok/ai_testing_common:__pkg__",
        "//src/test/cpp/net/eagle0/common/mcts:__subpackages__",
        "//src/test/cpp/net/eagle0/shardok/ai/mcts:__subpackages__",
    ],
@@ -54,7 +54,6 @@ struct MCTSNode {
    // Chance node specific fields (only used when nodeType == CHANCE)
    std::vector<double> outcomeProbabilities;  // Probability of each outcome
    std::vector<double> outcomeRolls;          // Representative roll for each outcome
-    double unvisitedOutcomeScore = 0.0;        // Default value for outcomes not expanded yet

    // Game context
    MCTSPlayerId playerId;
@@ -275,4 +274,4 @@ struct MCTSNode {
 }  // namespace mcts
 }  // namespace shardok

-#endif  // EAGLE0_ABSTRACT_MCTSNODE_HPP
+#endif  // EAGLE0_ABSTRACT_MCTSNODE_HPP
@@ -51,11 +51,10 @@ struct MCTSConfig {
                                     // When evaluating a leaf at playerFlips < maxSimulationFlips,
                                     // simulate forward to this phase for fair comparison
                                     // (default 0 = evaluate leaves as-is, backward compatible)
-    uint32_t randomSeed = 0;         // 0 = seed nondeterministically
    std::string debugDumpPath = "";  // If non-empty, dump MCTS tree to this file path
 };

 }  // namespace mcts
 }  // namespace shardok

-#endif  // EAGLE0_MCTS_TYPES_HPP
+#endif  // EAGLE0_MCTS_TYPES_HPP
@@ -54,12 +54,6 @@ static auto IsDeterministic(const CommandType type) -> bool {
    }
 }

-static auto RandomnessSampleForRepeat(const int repeatIteration, const int maxRepeatCount)
-        -> double {
-    if (maxRepeatCount <= 1) { return 0.5; }
-    return static_cast<double>(repeatIteration) / static_cast<double>(maxRepeatCount - 1);
-}
-
 // Helper function to sort commands by score
 static auto CommandSorter(
        const AICommandEvaluator::IndexAndScore& l,
@@ -91,15 +85,15 @@ auto AICommandEvaluator::PerformLookahead(
        const ScoreValue currentUtility,
        const AIStrategy& attackerStrategy,
        const CoordsSet& allCastleCoords,
-        std::chrono::steady_clock::time_point deadline) const -> std::future<EvaluationResult> {
+        std::chrono::steady_clock::time_point deadline) const -> std::future<ScoreValue> {
    // Check transposition table before expensive computation
    auto cachedScore =
            g_transpositionTable.probe(innerEngine->GetCurrentGameState(), remainingLookahead, pid);

    if (cachedScore.has_value()) {
        // Return cached result immediately
-        std::promise<EvaluationResult> p;
-        p.set_value(EvaluationResult{.score = *cachedScore, .completed = true});
+        std::promise<ScoreValue> p;
+        p.set_value(*cachedScore);
        return p.get_future();
    }
    const auto nextUtility = currentUtility;
@@ -111,8 +105,8 @@ auto AICommandEvaluator::PerformLookahead(
        // table
        g_transpositionTable.store(innerEngine->GetCurrentGameState(), 1, pid, nextUtility);

-        std::promise<EvaluationResult> p;
-        p.set_value(EvaluationResult{.score = nextUtility, .completed = true});
+        std::promise<ScoreValue> p;
+        p.set_value(nextUtility);
        return p.get_future();
    }

@@ -137,18 +131,16 @@ auto AICommandEvaluator::PerformLookahead(
                 innerEngine,
                 pid,
                 nextUtility,
-                 remainingLookahead]() mutable -> EvaluationResult {
-                    const auto bestCommand = bestCommandFuture.get();
-                    if (!bestCommand.completed) {
-                        return EvaluationResult{.score = nextUtility, .completed = false};
-                    }
+                 remainingLookahead]() mutable -> ScoreValue {
+                    const auto [index, type, lookaheadScore, immediateScore] =
+                            bestCommandFuture.get();

                    ScoreValue resultScore;
-                    if (auto& nextCommand = innerEngine->GetAvailableCommandsForAIPlayer(pid)->at(
-                                bestCommand.index);
+                    if (auto& nextCommand =
+                                innerEngine->GetAvailableCommandsForAIPlayer(pid)->at(index);
                        nextCommand->GetCommandType() !=
                        net::eagle0::shardok::common::END_TURN_COMMAND) {
-                        resultScore = bestCommand.immediateScore;
+                        resultScore = immediateScore;
                    } else {
                        resultScore = nextUtility;
                    }
@@ -160,7 +152,7 @@ auto AICommandEvaluator::PerformLookahead(
                            pid,
                            resultScore);

-                    return EvaluationResult{.score = resultScore, .completed = true};
+                    return resultScore;
                });
    }

@@ -168,8 +160,8 @@ auto AICommandEvaluator::PerformLookahead(
    g_transpositionTable
            .store(innerEngine->GetCurrentGameState(), remainingLookahead, pid, nextUtility);

-    std::promise<EvaluationResult> p;
-    p.set_value(EvaluationResult{.score = nextUtility, .completed = true});
+    std::promise<ScoreValue> p;
+    p.set_value(nextUtility);
    return p.get_future();
 }

@@ -188,10 +180,10 @@ auto AICommandEvaluator::EvaluateWithRandomness(

    // Check if we've exceeded the deadline
    if (std::chrono::steady_clock::now() > deadline) {
-        std::promise<EvaluationResult> p;
-        p.set_value(EvaluationResult{.score = 0.0, .completed = false});
+        // Return with a default score and an empty future that resolves immediately
+        std::promise<ScoreValue> p;
+        p.set_value(0.0);  // Default timeout score
        returnValue.immediateScore = 0.0;
-        returnValue.completed = false;
        returnValue.lookaheadScore = p.get_future();
        return returnValue;
    }
@@ -206,12 +198,11 @@ auto AICommandEvaluator::EvaluateWithRandomness(
            allCastleCoords);

    returnValue.immediateScore = innerUtility;
-    returnValue.completed = true;

    if (remainingLookahead <= 0) {
-        std::promise<EvaluationResult> p;
+        std::promise<ScoreValue> p;
        returnValue.lookaheadScore = p.get_future();
-        p.set_value(EvaluationResult{.score = innerUtility, .completed = true});
+        p.set_value(innerUtility);
    } else {
        auto lookaheadLambda = [this,
                                pid,
@@ -222,7 +213,7 @@ auto AICommandEvaluator::EvaluateWithRandomness(
                                attackerStrategy,
                                innerUtility,
                                &allCastleCoords,
-                                deadline]() -> EvaluationResult {
+                                deadline]() -> ScoreValue {
            auto lookaheadFuture = PerformLookahead(
                    pid,
                    isDefender,
@@ -240,7 +231,7 @@ auto AICommandEvaluator::EvaluateWithRandomness(
        auto launchPolicy = remainingLookahead == 1 ? std::launch::async : std::launch::deferred;
        returnValue.lookaheadScore = std::async(launchPolicy, lookaheadLambda);
 #else
-        std::promise<EvaluationResult> p;
+        std::promise<ScoreValue> p;
        returnValue.lookaheadScore = p.get_future();
        auto lambdaResult = lookaheadLambda();
        p.set_value(lambdaResult);
@@ -325,8 +316,7 @@ auto AICommandEvaluator::FindBestCommand(
        size_t index;
        CommandType type;
        ScoreValue immediateScore;
-        bool immediateCompleted = true;
-        std::vector<std::future<EvaluationResult>> lookaheadFutures;
+        std::vector<std::future<ScoreValue>> lookaheadFutures;
    };

    std::vector<CommandEvaluation> commandEvaluations(commandCount);
@@ -340,12 +330,12 @@ auto AICommandEvaluator::FindBestCommand(
        commandEvaluations[index].type = guessedCommandType;

        if (guessedCommandType == net::eagle0::shardok::common::END_TURN_COMMAND) {
-            std::promise<EvaluationResult> p;
+            std::promise<ScoreValue> p;
            commandEvaluations[index].lookaheadFutures.push_back(p.get_future());
-            p.set_value(EvaluationResult{.score = currentUtility, .completed = true});
+            p.set_value(currentUtility);
            commandEvaluations[index].immediateScore = currentUtility;
        } else if (IsDeterministic(guessedCommandType)) {
-            auto evaluation = EvaluateWithRandomness(
+            auto [immediateScore, lookaheadScore] = EvaluateWithRandomness(
                    pid,
                    isDefender,
                    originalIndex,
@@ -357,16 +347,14 @@ auto AICommandEvaluator::FindBestCommand(
                    allCastleCoords,
                    deadline);

-            commandEvaluations[index].immediateScore = evaluation.immediateScore;
-            commandEvaluations[index].immediateCompleted = evaluation.completed;
-            commandEvaluations[index].lookaheadFutures.push_back(
-                    std::move(evaluation.lookaheadScore));
+            commandEvaluations[index].immediateScore = immediateScore;
+            commandEvaluations[index].lookaheadFutures.push_back(std::move(lookaheadScore));
        } else if (guessedDescriptor->HasOdds()) {
            const auto successChancePercentile = guessedDescriptor->GetOddsPercentile();
            const double successChance = static_cast<double>(successChancePercentile) / 100.0;

            // Success attempt uses 1.0 - (successChance / 2) as the roll
-            auto successEvaluation = EvaluateWithRandomness(
+            auto [successImmediateScore, successLookaheadScore] = EvaluateWithRandomness(
                    pid,
                    isDefender,
                    originalIndex,
@@ -380,7 +368,7 @@ auto AICommandEvaluator::FindBestCommand(
                    deadline);

            // Failure attempt uses the average of (1 - successChance) and 0 as the roll
-            auto failureEvaluation = EvaluateWithRandomness(
+            auto [failureImmediateScore, failureLookaheadScore] = EvaluateWithRandomness(
                    pid,
                    isDefender,
                    originalIndex,
@@ -393,35 +381,24 @@ auto AICommandEvaluator::FindBestCommand(
                    allCastleCoords,
                    deadline);

-            commandEvaluations[index].immediateCompleted =
-                    failureEvaluation.completed && successEvaluation.completed;
-            commandEvaluations[index].immediateScore = std::lerp(
-                    failureEvaluation.immediateScore,
-                    successEvaluation.immediateScore,
-                    successChance);
+            commandEvaluations[index].immediateScore =
+                    std::lerp(failureImmediateScore, successImmediateScore, successChance);

-            auto successSF = successEvaluation.lookaheadScore.share();
-            auto failureSF = failureEvaluation.lookaheadScore.share();
+            auto successSF = successLookaheadScore.share();
+            auto failureSF = failureLookaheadScore.share();
            commandEvaluations[index].lookaheadFutures.push_back(std::async(
                    std::launch::deferred,
-                    [successSF, failureSF, successChance]() -> EvaluationResult {
-                        const auto failure = failureSF.get();
-                        const auto success = successSF.get();
-                        if (!failure.completed || !success.completed) {
-                            return EvaluationResult{.score = 0.0, .completed = false};
-                        }
-                        return EvaluationResult{
-                                .score = std::lerp(failure.score, success.score, successChance),
-                                .completed = true};
+                    [successSF, failureSF, successChance]() -> double {
+                        return std::lerp(failureSF.get(), successSF.get(), successChance);
                    }));
        } else {
            ScoreValue sum = 0.0;
-            const int sampleCount = std::max(1, maxRepeatCount);
-            for (int repeatIteration = 0; repeatIteration < sampleCount; repeatIteration++) {
+            for (int repeatIteration = 0; repeatIteration < maxRepeatCount; repeatIteration++) {
                // In each iteration, use a double from [0, 1] as the random roll
-                auto sequence =
-                        std::vector{RandomnessSampleForRepeat(repeatIteration, sampleCount)};
-                auto evaluation = EvaluateWithRandomness(
+                auto sequence = std::vector{
+                        static_cast<double>(repeatIteration) /
+                        static_cast<double>(maxRepeatCount - 1)};
+                auto [immediateScore, lookaheadScore] = EvaluateWithRandomness(
                        pid,
                        isDefender,
                        originalIndex,
@@ -433,12 +410,10 @@ auto AICommandEvaluator::FindBestCommand(
                        allCastleCoords,
                        deadline);

-                if (!evaluation.completed) { commandEvaluations[index].immediateCompleted = false; }
-                sum += evaluation.immediateScore;
-                commandEvaluations[index].lookaheadFutures.push_back(
-                        std::move(evaluation.lookaheadScore));
+                sum += immediateScore;
+                commandEvaluations[index].lookaheadFutures.push_back(std::move(lookaheadScore));
            }
-            commandEvaluations[index].immediateScore = sum / sampleCount;
+            commandEvaluations[index].immediateScore = sum / maxRepeatCount;
        }
    }

@@ -452,33 +427,19 @@ auto AICommandEvaluator::FindBestCommand(
                // Wait for all futures and compute final scores
                for (auto& eval : evals) {
                    ScoreValue totalLookaheadScore = 0.0;
-                    bool completed = eval.immediateCompleted;
                    for (auto& future : eval.lookaheadFutures) {
-                        const auto lookahead = future.get();
-                        if (!lookahead.completed) { completed = false; }
-                        totalLookaheadScore += lookahead.score;
+                        totalLookaheadScore += future.get();
                    }
                    ScoreValue avgLookaheadScore =
                            eval.lookaheadFutures.empty()
                                    ? eval.immediateScore
                                    : totalLookaheadScore / eval.lookaheadFutures.size();

-                    if (!completed) { continue; }
-
                    allResults.push_back(IndexAndScore{
                            .index = eval.index,
                            .type = eval.type,
                            .lookaheadScore = avgLookaheadScore,
-                            .immediateScore = eval.immediateScore,
-                            .completed = true});
-                }
-                if (allResults.empty()) {
-                    return IndexAndScore{
-                            .index = 0,
-                            .type = net::eagle0::shardok::common::UNKNOWN_COMMAND,
-                            .lookaheadScore = 0.0,
-                            .immediateScore = 0.0,
-                            .completed = false};
+                            .immediateScore = eval.immediateScore});
                }
                // Find the best command using the existing sorter
                auto bestIt = std::ranges::max_element(allResults, CommandSorter);
@@ -496,12 +457,12 @@ auto AICommandEvaluator::EvaluateCommand(
        const ScoreValue currentUtility,
        const CoordsSet& allCastleCoords,
        const size_t commandIndex,
-        std::chrono::steady_clock::time_point deadline) const -> std::future<EvaluationResult> {
+        std::chrono::steady_clock::time_point deadline) const -> std::future<ScoreValue> {
    const CommandListSPtr guessedDescriptors = guessedEngine.GetAvailableCommandsForAIPlayer(pid);

    if (commandIndex >= guessedDescriptors->size()) {
-        std::promise<EvaluationResult> p;
-        p.set_value(EvaluationResult{.score = currentUtility, .completed = true});
+        std::promise<ScoreValue> p;
+        p.set_value(currentUtility);
        return p.get_future();
    }

@@ -509,11 +470,11 @@ auto AICommandEvaluator::EvaluateCommand(

    if (const auto guessedCommandType = guessedDescriptor->GetCommandType();
        guessedCommandType == net::eagle0::shardok::common::END_TURN_COMMAND) {
-        std::promise<EvaluationResult> p;
-        p.set_value(EvaluationResult{.score = currentUtility, .completed = true});
+        std::promise<ScoreValue> p;
+        p.set_value(currentUtility);
        return p.get_future();
    } else if (IsDeterministic(guessedCommandType)) {
-        auto evaluation = EvaluateWithRandomness(
+        auto [immediateScore, lookaheadScore] = EvaluateWithRandomness(
                pid,
                isDefender,
                commandIndex,
@@ -524,13 +485,13 @@ auto AICommandEvaluator::EvaluateCommand(
                attackerStrategy,
                allCastleCoords,
                deadline);
-        return std::move(evaluation.lookaheadScore);
+        return std::move(lookaheadScore);
    } else if (guessedDescriptor->HasOdds()) {
        const auto successChancePercentile = guessedDescriptor->GetOddsPercentile();
        const double successChance = static_cast<double>(successChancePercentile) / 100.0;

        // Success attempt
-        auto successEvaluation = EvaluateWithRandomness(
+        auto [successImmediateScore, successLookaheadScore] = EvaluateWithRandomness(
                pid,
                isDefender,
                commandIndex,
@@ -543,7 +504,7 @@ auto AICommandEvaluator::EvaluateCommand(
                deadline);

        // Failure attempt
-        auto failureEvaluation = EvaluateWithRandomness(
+        auto [failureImmediateScore, failureLookaheadScore] = EvaluateWithRandomness(
                pid,
                isDefender,
                commandIndex,
@@ -556,29 +517,20 @@ auto AICommandEvaluator::EvaluateCommand(
                deadline);

        // Return weighted average of success and failure
-        auto successSF = successEvaluation.lookaheadScore.share();
-        auto failureSF = failureEvaluation.lookaheadScore.share();
-        return std::async(
-                std::launch::deferred,
-                [successSF, failureSF, successChance]() -> EvaluationResult {
-                    const auto failure = failureSF.get();
-                    const auto success = successSF.get();
-                    if (!failure.completed || !success.completed) {
-                        return EvaluationResult{.score = 0.0, .completed = false};
-                    }
-                    return EvaluationResult{
-                            .score = std::lerp(failure.score, success.score, successChance),
-                            .completed = true};
-                });
+        auto successSF = successLookaheadScore.share();
+        auto failureSF = failureLookaheadScore.share();
+        return std::async(std::launch::deferred, [successSF, failureSF, successChance]() -> double {
+            return std::lerp(failureSF.get(), successSF.get(), successChance);
+        });
    } else {
        // For non-deterministic commands without odds, use multiple attempts
-        std::vector<std::future<EvaluationResult>> lookaheadFutures;
-        const int sampleCount = std::max(1, maxRepeatCount);
-        lookaheadFutures.reserve(sampleCount);
+        std::vector<std::future<ScoreValue>> lookaheadFutures;
+        lookaheadFutures.reserve(maxRepeatCount);

-        for (int repeatIteration = 0; repeatIteration < sampleCount; repeatIteration++) {
-            auto sequence = std::vector{RandomnessSampleForRepeat(repeatIteration, sampleCount)};
-            auto evaluation = EvaluateWithRandomness(
+        for (int repeatIteration = 0; repeatIteration < maxRepeatCount; repeatIteration++) {
+            auto sequence = std::vector{
+                    static_cast<double>(repeatIteration) / static_cast<double>(maxRepeatCount - 1)};
+            auto [immediateScore, lookaheadScore] = EvaluateWithRandomness(
                    pid,
                    isDefender,
                    commandIndex,
@@ -590,25 +542,17 @@ auto AICommandEvaluator::EvaluateCommand(
                    allCastleCoords,
                    deadline);

-            lookaheadFutures.push_back(std::move(evaluation.lookaheadScore));
+            lookaheadFutures.push_back(std::move(lookaheadScore));
        }

        // Return a future that computes the average when needed
        return std::async(
                std::launch::deferred,
                [lookaheadFutures = std::move(lookaheadFutures),
-                 sampleCount]() mutable -> EvaluationResult {
+                 maxRepeatCount]() mutable -> double {
                    ScoreValue total = 0.0;
-                    bool completed = true;
-                    for (auto& future : lookaheadFutures) {
-                        const auto result = future.get();
-                        if (!result.completed) { completed = false; }
-                        total += result.score;
-                    }
-                    if (!completed) { return EvaluationResult{.score = 0.0, .completed = false}; }
-                    return EvaluationResult{
-                            .score = total / static_cast<ScoreValue>(sampleCount),
-                            .completed = true};
+                    for (auto& future : lookaheadFutures) { total += future.get(); }
+                    return total / maxRepeatCount;
                });
    }
 }
@@ -38,11 +38,6 @@ public:
            BattalionTypeGetter battalionTypeGetter);  // Pass by value

    /// Evaluates the score for a particular command index with lookahead.
-    struct EvaluationResult {
-        ScoreValue score;
-        bool completed;
-    };
-
    [[nodiscard]] auto EvaluateCommand(
            PlayerId pid,
            bool isDefender,
@@ -53,7 +48,7 @@ public:
            ScoreValue currentUtility,
            const CoordsSet& allCastleCoords,
            size_t commandIndex,
-            std::chrono::steady_clock::time_point deadline) const -> std::future<EvaluationResult>;
+            std::chrono::steady_clock::time_point deadline) const -> std::future<ScoreValue>;

    /// Find the best command among all available commands at the given depth.
    struct IndexAndScore {
@@ -61,7 +56,6 @@ public:
        CommandType type;
        ScoreValue lookaheadScore;
        ScoreValue immediateScore;
-        bool completed;
    };

    [[nodiscard]] auto FindBestCommand(
@@ -82,8 +76,7 @@ private:

    struct ImmediateAndLookaheadScore {
        ScoreValue immediateScore;
-        bool completed;
-        std::future<EvaluationResult> lookaheadScore;
+        std::future<ScoreValue> lookaheadScore;
    };

    /// Recursive lookahead calculator
@@ -96,7 +89,7 @@ private:
            ScoreValue currentUtility,
            const AIStrategy& attackerStrategy,
            const CoordsSet& allCastleCoords,
-            std::chrono::steady_clock::time_point deadline) const -> std::future<EvaluationResult>;
+            std::chrono::steady_clock::time_point deadline) const -> std::future<ScoreValue>;

    /// Evaluate single command execution with randomness handling
    [[nodiscard]] auto EvaluateWithRandomness(
@@ -75,47 +75,15 @@ std::vector<size_t> AICommandFilter::FilterCommands(
    // Calculate minimum distance to enemies once for all filters
    const double minDistToEnemies = MinDistanceToEnemyUnits(gameState, pid, enemyLocations);

-    bool hasAvailablePlacementCommand = false;
-    for (const auto& cmd : *commands) {
-        if (cmd->GetCommandType() == CommandType::PLACE_UNIT_COMMAND ||
-            cmd->GetCommandType() == CommandType::PLACE_HIDDEN_UNIT_COMMAND) {
-            hasAvailablePlacementCommand = true;
-            break;
-        }
-    }
-
-    bool playerHasReserveUnits = false;
-    for (const Unit* unit : *gameState->units()) {
-        if (unit->player_id() == pid &&
-            unit->status() == net::eagle0::shardok::storage::fb::UnitStatus_RESERVE_UNIT) {
-            playerHasReserveUnits = true;
-            break;
-        }
-    }
-
    for (size_t i = 0; i < commands->size(); ++i) {
        const auto& cmd = (*commands)[i];

-        if (cmd->GetCommandType() == CommandType::END_PLAYER_SETUP_COMMAND &&
-            hasAvailablePlacementCommand && playerHasReserveUnits) {
-            continue;
-        }
-
        // Always allow END_TURN commands
        if (cmd->GetCommandType() == CommandType::END_TURN_COMMAND) {
            filteredIndices.push_back(i);
            continue;
        }

-        // Setup placement commands should always remain available. The movement/spell heuristics
-        // below are combat-oriented and can otherwise filter every placement for an in-reserve
-        // unit, making MCTS think ending setup is the only good option.
-        if (cmd->GetCommandType() == CommandType::PLACE_UNIT_COMMAND ||
-            cmd->GetCommandType() == CommandType::PLACE_HIDDEN_UNIT_COMMAND) {
-            filteredIndices.push_back(i);
-            continue;
-        }
-
        // Filter obviously bad moves
        bool shouldFilter = false;

@@ -700,4 +668,4 @@ bool AICommandFilter::WouldAbandonCriticalCastle(
    return false;
 }

-}  // namespace shardok
+}  // namespace shardok
@@ -15,11 +15,11 @@ enum class AIAlgorithmType {

 // Enum for scoring calculator selection
 enum class ScoringCalculatorType {
-    STANDARD,        // Default: Unbounded raw scores
-    MCTS_OPTIMIZED,  // Bounded linear scores tuned for MCTS
-    EXPERIMENTAL     // Standard-derived scorer selected by experiment id
+    STANDARD,       // Default: Unbounded raw scores
+    NORMALIZED,     // Normalized scores in [0, 1] range for ML training
+    MCTS_OPTIMIZED  // Bounded linear scores tuned for MCTS
 };

 }  // namespace shardok

-#endif  // EAGLE0_AI_CONFIG_HPP
+#endif  // EAGLE0_AI_CONFIG_HPP
@@ -84,7 +84,6 @@ auto AIDefenderStrategySelector::BestDefenderStrategy(
    //  * the defender has way, way fewer troops
    // other considerations we should have but do not presently:
    //  * the attacker is close to the defender / castles
-
    if (canFlee && roundsRemaining > 2 &&
        defenderTroops < MAXIMUM_RATIO_FOR_DEFENDER_TO_FLEE * attackerTroops &&
        attackerNonUndeadUnitNotRequiringWaterCrossingCount >= criticalTileCoords.size()) {
@@ -1,23 +0,0 @@
-//
-// Temporary AI experiment switches used by benchmark-only EXPERIMENTAL runs.
-//
-
-#ifndef EAGLE0_AI_EXPERIMENT_CONFIG_HPP
-#define EAGLE0_AI_EXPERIMENT_CONFIG_HPP
-
-#include <cstdlib>
-#include <string>
-
-namespace shardok {
-
-inline auto CurrentAIExperimentId() -> int {
-    const char* rawValue = std::getenv("SHARDOK_SCORING_EXPERIMENT_ID");
-    if (rawValue == nullptr || std::string(rawValue).empty()) { return 0; }
-    return std::atoi(rawValue);
-}
-
-inline auto IsAIExperiment(const int id) -> bool { return CurrentAIExperimentId() == id; }
-
-}  // namespace shardok
-
-#endif  // EAGLE0_AI_EXPERIMENT_CONFIG_HPP
@@ -5,6 +5,7 @@
 #include "AIUnitScoreCalculator.hpp"

 #include <algorithm>
+#include <cstdlib>

 #include "AIAttackLocations.hpp"
 #include "src/main/cpp/net/eagle0/shardok/library/util/HexMapUtils.hpp"
@@ -32,9 +33,7 @@ constexpr double kMeteorCastRoundMultiplier = 1.5;
 constexpr double kMeteorCastInsufficientVigorMultiplier = 0.5;
 constexpr double kMeteorCastleMultiplier = 2.0;
 constexpr double kLightningPossibleValue = 0.05;
-constexpr double kExpectedArcheryVolleyTargetValueFraction = 0.16;
-constexpr double kImmediateArcheryThreatMultiplier = 0.5;
-constexpr double kFutureArcheryThreatMultiplier = 0.25;
+constexpr double kArcheryPossibleValue = 38;
 constexpr double kReduceFortifiedPossibleValue = 50;
 constexpr double kReduceUnfortifiedPossibleValue = 12;
 constexpr double kFearPossibleValue = 38;
@@ -102,24 +101,9 @@ auto ContextFreeUnitValue(const Unit *unit) -> ScoreValue {
    return battalionValue + heroValue;
 }

-auto bestEnemyUnitValue(const vector<const Unit *> &enemyUnits) -> double {
-    double bestTargetValue = 0.0;
-    for (const Unit *enemyUnit : enemyUnits) {
-        bestTargetValue = std::max(bestTargetValue, ContextFreeUnitValue(enemyUnit));
-    }
-    return bestTargetValue;
-}
-
-auto expectedImpactArcheryValue(const vector<const Unit *> &enemyUnits, const bool canShootNow)
-        -> double {
-    const double expectedVolleyImpact =
-            kExpectedArcheryVolleyTargetValueFraction * bestEnemyUnitValue(enemyUnits);
-    return expectedVolleyImpact *
-           (canShootNow ? kImmediateArcheryThreatMultiplier : kFutureArcheryThreatMultiplier);
-}
-
-auto archeryValue(const vector<const Unit *> &enemyUnits, const bool canShootNow) -> double {
-    return expectedImpactArcheryValue(enemyUnits, canShootNow);
+auto archeryValue(const Unit * /*unit*/) -> double {
+    // TODO: make this depend on the value of the targets
+    return kArcheryPossibleValue;
 }

 auto reduceValue(const Unit *unit, const Terrain *unitTerrain) -> double {
@@ -132,7 +116,10 @@ auto reduceValue(const Unit *unit, const Terrain *unitTerrain) -> double {
        return 0.0;
 }

-auto fearValue(const vector<const Unit *> &) -> double { return kFearPossibleValue; }
+auto fearValue(const Unit * /*unit*/) -> double {
+    // TODO: make this depend on the value of the targets
+    return kFearPossibleValue;
+}

 auto lightningValue(const Unit *unit) -> double {
    // TODO: make this depend on the value of the targets
@@ -278,13 +265,11 @@ auto GetRangedAttackBonus(
        const vector<const Unit *> &attackerUnits,
        const vector<const Unit *> &defenderUnits,
        const int meteorRange,
-        const double minVigorToCast,
-        const ActionPoints archeryActionPointCost) -> double {
+        const double minVigorToCast) -> double {
    vector<double> rangedAttackValues{};

    const auto &enemyUnits = isAttacker ? defenderUnits : attackerUnits;
    const auto &friendlyUnits = isAttacker ? attackerUnits : defenderUnits;
-    const bool canShootNow = unit->remaining_action_points() >= archeryActionPointCost;

    const auto &unitLocation = unit->location();

@@ -321,19 +306,19 @@ auto GetRangedAttackBonus(
        unit->attached_hero().profession_info().profession() ==
                net::eagle0::shardok::storage::fb::Profession_NECROMANCER &&
        unit->attached_hero().control_info().controlled_unit_id() != -1) {
-        rangedAttackValues.push_back(fearValue(enemyUnits));
+        rangedAttackValues.push_back(fearValue(unit));
    }

    if (attackLocations.ArcheryLocations().Contains(unitLocation) &&
        unit->volleys_remaining() > 0) {
-        rangedAttackValues.push_back(archeryValue(enemyUnits, canShootNow));
+        rangedAttackValues.push_back(archeryValue(unit));
    }

    // Longbowmen in castle can archer adjacent
    if (attackLocations.AdjacentLocations().Contains(unitLocation) &&
        unit->volleys_remaining() > 0 && terrain->modifier().castle().present() &&
        unit->battalion().type() == net::eagle0::shardok::storage::fb::BattalionTypeId_LONGBOWMEN) {
-        rangedAttackValues.push_back(archeryValue(enemyUnits, canShootNow));
+        rangedAttackValues.push_back(archeryValue(unit));
    }

    if (rangedAttackValues.empty()) return 0.0;
@@ -353,8 +338,7 @@ auto UnitValue(
        const CoordsSet &locationsInDangerFromEnemy,
        const ActionPointDistances *distances,
        int meteorRange,
-        double meteorCastVigorCost,
-        ActionPoints archeryActionPointCost) -> ScoreValue {
+        double meteorCastVigorCost) -> ScoreValue {
    const auto &location = unit->location();
    if (location.row() < 0) return 0;  // unplaced unit

@@ -398,8 +382,7 @@ auto UnitValue(
            attackerUnits,
            defenderUnits,
            meteorRange,
-            meteorCastVigorCost,
-            archeryActionPointCost);
+            meteorCastVigorCost);

    // scouting values
    // attack range
@@ -32,8 +32,7 @@ auto GetRangedAttackBonus(
        const vector<const Unit *> &attackerUnits,
        const vector<const Unit *> &defenderUnits,
        int meteorRange,
-        double minVigorToCast,
-        ActionPoints archeryActionPointCost = 0) -> double;
+        double minVigorToCast) -> double;

 auto UnitValue(
        const Unit *unit,
@@ -48,8 +47,7 @@ auto UnitValue(
        const CoordsSet &locationsInDangerFromEnemy,
        const ActionPointDistances *distances,
        int meteorRange,
-        double meteorCastVigorCost,
-        ActionPoints archeryActionPointCost = 0) -> ScoreValue;
+        double meteorCastVigorCost) -> ScoreValue;

 }  // namespace shardok

@@ -156,9 +156,7 @@ The system uses recursive lookahead with:
 - `kProfessionValue`: 200

 ### Ranged Attack Values
- Archery position value: expected-impact based. A unit that can shoot now gets half of
-  `0.16 * best target unit value`; a unit in archery range that cannot shoot yet gets one
-  quarter of that expected volley value.
+- `kArcheryPossibleValue`: 38
 - `kMeteorDirectTargetingEnemy`: 2 per soldier
 - `kMeteorSplashTargetingEnemy`: 1 per soldier
 - `kLightningPossibleValue`: 0.05 per soldier
@@ -764,4 +762,4 @@ MCTSAI ai(playerId, isDefender, strategy, castleCoords, apdCache, alCache, confi
 - Higher `immediateScoreTieBreakThreshold` to emphasize direct paths
 - `BEST_IMMEDIATE` simulation for most predictable behavior

-The configuration system allows fine-tuning MCTS behavior for different scenarios while maintaining compatibility with the existing AI infrastructure.
+The configuration system allows fine-tuning MCTS behavior for different scenarios while maintaining compatibility with the existing AI infrastructure.
@@ -11,13 +11,6 @@ cc_library(
    ],
 )

-cc_library(
-    name = "ai_experiment_config",
-    hdrs = ["AIExperimentConfig.hpp"],
-    copts = COPTS,
-    visibility = ["//src/main/cpp/net/eagle0/shardok/ai:__subpackages__"],
-)
-
 cc_library(
    name = "ai_attacker_strategy_selector",
    srcs = ["AIAttackerStrategySelector.cpp"],
@@ -91,7 +84,6 @@ cc_library(
    ],
    deps = [
        ":ai_attack_locations",
-        ":ai_experiment_config",
        ":ai_score_utilities",
        ":ai_strategy",
        ":ai_water_crossing_calculator",
@@ -204,7 +196,6 @@ cc_library(
    ],
    deps = [
        ":ai_command_filter",
-        ":ai_experiment_config",
        ":ai_strategy",
        ":transposition_table",
        "//src/main/cpp/net/eagle0/common:sequence_random_generator",
@@ -281,7 +272,6 @@ cc_library(
    ],
    deps = [
        ":ai_attack_locations",
-        ":ai_experiment_config",
        "//src/main/cpp/net/eagle0/shardok/library/action_point_distances",
        "//src/main/cpp/net/eagle0/shardok/library/util:hex_map_utils",
    ],
@@ -358,7 +348,6 @@ cc_library(
        ":ai_attacker_strategy_selector",
        ":ai_command_evaluator",
        ":ai_defender_strategy_selector",
-        ":ai_experiment_config",
        ":ai_time_budget",
        ":ai_water_crossing_command_chooser",
        "//src/main/cpp/net/eagle0/common:time_utils",
@@ -396,6 +385,7 @@ cc_library(
        "//src/main/cpp/net/eagle0/shardok/ai/mcts:shardok_mcts_ai",  # MCTS with abstraction layer
        "//src/main/cpp/net/eagle0/shardok/ai/score:ai_score_calculator_interface",
        "//src/main/cpp/net/eagle0/shardok/ai/score:mcts_optimized_ai_score_calculator",  # Bounded linear scorer for MCTS
+        "//src/main/cpp/net/eagle0/shardok/ai/score:normalized_ai_score_calculator",  # Normalized [0,1] scorer for ML training
        "//src/main/cpp/net/eagle0/shardok/ai/score:standard_ai_score_calculator",  # Standard unbounded scorer (default)
        "//src/main/cpp/net/eagle0/shardok/library/action_point_distances",
        "//src/main/cpp/net/eagle0/shardok/library/action_point_distances:action_point_distances_cache",
@@ -13,7 +13,6 @@
 #include "AICommandEvaluator.hpp"
 #include "AICommandFilter.hpp"
 #include "TranspositionTable.hpp"
-#include "src/main/cpp/net/eagle0/shardok/ai/AIExperimentConfig.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai/score/AIScoreCalculator.hpp"
 #include "src/main/cpp/net/eagle0/shardok/library/ShardokEngine.hpp"

@@ -131,10 +130,6 @@ auto IterativeDeepeningAI::IterativeSearch(
        // Now wait for all futures and collect results
        for (auto& [cmdIndex, future] : futures) {
            auto cmdResult = future.get();
-            if (!cmdResult.searchCompleted) {
-                allEvaluated = false;
-                continue;
-            }

            // Ensure scoresByDepth[cmdIndex] has enough space
            if (scoresByDepth[cmdIndex].size() <= currentDepth) {
@@ -226,7 +221,7 @@ auto IterativeDeepeningAI::IterativeSearch(
            }

            // If all evaluated commands had unchanged scores, we've hit END_TURN in lookahead
-            if (!IsAIExperiment(26) && scoresUnchanged && unchangedCount == evaluatedCount) {
+            if (scoresUnchanged && unchangedCount == evaluatedCount) {
                completionReason = EvaluationCompletionReason::RAN_OUT_OF_COMMANDS;
                break;
            }
@@ -256,9 +251,7 @@ auto IterativeDeepeningAI::IterativeSearch(
        completionReason = EvaluationCompletionReason::RAN_OUT_OF_COMMANDS;
    }

-    // Select best result from the highest completed depth achieved for each command. Commands that
-    // timed out at a depth are not recorded for that depth, so partial deeper work can help without
-    // letting timeout fallback scores compete.
+    // Select best result from highest depth achieved for each command
    result = SelectBestResult(scoresByDepth, highestDepthCompleted);
    result.minimumDepthCompleted = result.depthAchieved >= timeBudget.minDepthRequired;
    result.searchCompleted = result.minimumDepthCompleted;
@@ -349,9 +342,7 @@ auto IterativeDeepeningAI::SearchCommandAtDepthWithEngine(
    // Deduct adjusted time from remaining budget
    timeBudget.remainingBudget -= adjustedElapsedMs;

-    result.bestScore = commandScore.score;
-    result.searchCompleted = commandScore.completed;
-    result.minimumDepthCompleted = commandScore.completed;
+    result.bestScore = commandScore;

    std::promise<SearchResult> p;
    p.set_value(result);
@@ -364,7 +355,7 @@ auto IterativeDeepeningAI::GetCommandsSortedByPreviousDepth(
        const std::vector<size_t>& highestDepthCompleted,
        const std::vector<size_t>& filteredIndices) -> std::vector<size_t> {
    if (currentDepth == 1) {
-        // For depth 1, preserve the upstream command-factory priority order.
+        // For depth 1, return filtered indices in natural order
        return filteredIndices;
    }

@@ -401,20 +392,10 @@ auto IterativeDeepeningAI::SelectBestResult(
    result.bestScore = -std::numeric_limits<ScoreValue>::infinity();
    result.searchCompleted = false;

-    size_t selectionDepth = 0;
-    if (IsAIExperiment(24)) {
-        selectionDepth = std::numeric_limits<size_t>::max();
-        for (const size_t depth : highestDepthCompleted) {
-            if (depth > 0) { selectionDepth = std::min(selectionDepth, depth); }
-        }
-        if (selectionDepth == std::numeric_limits<size_t>::max()) { selectionDepth = 0; }
-    }
-
    // Find the command with best score at its highest evaluated depth
    for (size_t i = 0; i < scoresByDepth.size(); ++i) {
        if (highestDepthCompleted[i] > 0) {
-            const size_t depth = selectionDepth > 0 ? selectionDepth : highestDepthCompleted[i];
-            if (highestDepthCompleted[i] < depth || scoresByDepth[i].size() <= depth) { continue; }
+            const size_t depth = highestDepthCompleted[i];
            if (ScoreValue score = scoresByDepth[i][depth]; score > result.bestScore) {
                result.bestScore = score;
                result.bestCommandIndex = i;
@@ -426,4 +407,4 @@ auto IterativeDeepeningAI::SelectBestResult(
    return result;
 }

-}  // namespace shardok
+}  // namespace shardok
@@ -111,4 +111,4 @@ private:

 }  // namespace shardok

-#endif  // EAGLE0_ITERATIVEDEEPENINGAI_HPP
+#endif  // EAGLE0_ITERATIVEDEEPENINGAI_HPP
@@ -32,6 +32,7 @@
 #include "IterativeDeepeningAI.hpp"
 #include "mcts/ShardokMCTSAI.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai/score/MCTSOptimizedAIScoreCalculator.hpp"
+#include "src/main/cpp/net/eagle0/shardok/ai/score/NormalizedAIScoreCalculator.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai/score/StandardAIScoreCalculator.hpp"
 #include "src/main/cpp/net/eagle0/shardok/library/util/HexMapUtils.hpp"
 #include "src/main/cpp/net/eagle0/shardok/library/view_filters/GameStateGuesser.hpp"
@@ -298,12 +299,12 @@ auto ShardokAIClient::StandardChooseCommandIndex(
    // Create scorer for actual scoring during search - type selected at construction
    std::unique_ptr<AIScoreCalculator> scorer;
    switch (scoringCalculatorType) {
+        case ScoringCalculatorType::NORMALIZED:
+            scorer = MakeNormalizedAIScoreCalculator(settingsGetter, apdCache, alCache);
+            break;
        case ScoringCalculatorType::MCTS_OPTIMIZED:
            scorer = MakeMCTSOptimizedAIScoreCalculator(settingsGetter, apdCache, alCache);
            break;
-        case ScoringCalculatorType::EXPERIMENTAL:
-            scorer = MakeExperimentalAIScoreCalculator(settingsGetter, apdCache, alCache);
-            break;
        case ScoringCalculatorType::STANDARD:
        default: scorer = MakeStandardAIScoreCalculator(settingsGetter, apdCache, alCache); break;
    }
@@ -21,22 +21,6 @@

 namespace shardok::mcts {

-namespace {
-
-auto IsDefenderPlayer(const GameStateW& gameState, const PlayerId playerId) -> bool {
-    for (const auto* pi : *gameState->player_infos()) {
-        if (pi->player_id() == playerId) { return pi->is_defender(); }
-    }
-    return false;
-}
-
-auto LegalActionsCacheKey(const uint64_t stateHash, const bool rootIsDefender) -> uint64_t {
-    constexpr uint64_t kDefenderRoleHash = 0x9e3779b97f4a7c15ULL;
-    return rootIsDefender ? stateHash ^ kDefenderRoleHash : stateHash;
-}
-
-}  // namespace
-
 // Shared cache for legal actions (uses lock-free parallel hash map for thread safety)
 // Using 8 submaps to reduce contention with 16 MCTS threads
 gtl::parallel_flat_hash_map<
@@ -239,14 +223,13 @@ std::vector<std::unique_ptr<MCTSAction>> ShardokGameEngine::getLegalActions(
    // Time hash computation
    const auto hashStart = std::chrono::high_resolution_clock::now();
    const uint64_t stateHash = shardokState->hash();
-    const uint64_t cacheKey = LegalActionsCacheKey(stateHash, isDefender_);
    const auto hashEnd = std::chrono::high_resolution_clock::now();
    timeInHashComputation_.fetch_add(
            std::chrono::duration_cast<std::chrono::microseconds>(hashEnd - hashStart).count(),
            std::memory_order_relaxed);

    // Check transposition table for cached legal actions
-    if (auto it = legalActionsCache_.find(cacheKey); it != legalActionsCache_.end()) {
+    if (auto it = legalActionsCache_.find(stateHash); it != legalActionsCache_.end()) {
        cacheHits_.fetch_add(1, std::memory_order_relaxed);

        // Use cached engine
@@ -321,7 +304,7 @@ std::vector<std::unique_ptr<MCTSAction>> ShardokGameEngine::getLegalActions(
    const std::vector<size_t> filteredIndices = AICommandFilter::FilterCommands(
            commands,
            currentPlayer,
-            IsDefenderPlayer(shardokState->getShardokState(), currentPlayer),
+            isDefender_,
            shardokState->getShardokState(),
            *apdCache_,
            [this](BattalionTypeId typeId) {
@@ -379,7 +362,7 @@ std::vector<std::unique_ptr<MCTSAction>> ShardokGameEngine::getLegalActions(
    // This avoids duplicating heavy protocol buffer objects
    // Use lazy_emplace_l to ensure thread-safe insertion (locks the bucket during construction)
    legalActionsCache_.lazy_emplace_l(
-            cacheKey,
+            stateHash,
            [&](typename decltype(legalActionsCache_)::value_type& v) {
                // Update existing entry
                v.second.filteredIndices = filteredIndices;
@@ -387,7 +370,7 @@ std::vector<std::unique_ptr<MCTSAction>> ShardokGameEngine::getLegalActions(
            },
            [&](const typename decltype(legalActionsCache_)::constructor& ctor) {
                // Create new entry
-                ctor(cacheKey, LegalActionsCache{filteredIndices, engine});
+                ctor(stateHash, LegalActionsCache{filteredIndices, engine});
            });

    return actions;
@@ -433,8 +416,14 @@ std::vector<double> ShardokGameEngine::getActionWeights(

    // Determine if current player is defender (not root player!)
    // During simulation we need to use the correct perspective for action weighting
+    bool currentPlayerIsDefender = false;
    const auto& gameState = shardokState->getShardokState();
-    const bool currentPlayerIsDefender = IsDefenderPlayer(gameState, currentPlayer);
+    for (const auto* pi : *gameState->player_infos()) {
+        if (pi->player_id() == currentPlayer) {
+            currentPlayerIsDefender = pi->is_defender();
+            break;
+        }
+    }

    // Use AIHeuristicWeighting for fast O(1) context-aware command weighting
    std::vector<double> weights;
@@ -110,19 +110,8 @@ bool ShardokGameState::equals(const MCTSGameState& other) const {
 }

 MCTSPlayerId ShardokGameState::getWinner() const {
-    if (!state_->status() ||
-        state_->status()->state() !=
-                net::eagle0::shardok::storage::fb::GameStatus_::State_VICTORY) {
-        return -1;
-    }
-
-    const auto* winningIds = state_->status()->winning_shardok_ids();
-    if (!winningIds || winningIds->empty()) { return -1; }
-
-    for (const PlayerId winningPid : *winningIds) {
-        if (winningPid >= 0) { return winningPid; }
-    }
-
+    // Note: FlatBuffer doesn't have a winner field
+    // In practice, this would need to determine winner from victory conditions
    return -1;  // No winner
 }

@@ -41,6 +41,31 @@ cc_library(
    ],
 )

+cc_library(
+    name = "normalized_ai_score_calculator",
+    srcs = ["NormalizedAIScoreCalculator.cpp"],
+    hdrs = ["NormalizedAIScoreCalculator.hpp"],
+    copts = COPTS,
+    visibility = [
+        "//src/main/cpp/net/eagle0/shardok/ai:__subpackages__",
+        "//src/main/cpp/net/eagle0/shardok/ai_performance_runner:__pkg__",
+        "//src/test/cpp/net/eagle0/shardok/ai:__subpackages__",
+    ],
+    deps = [
+        ":ai_score_calculator_interface",
+        ":ai_victory_condition_score_calculator",
+        "//src/main/cpp/net/eagle0/shardok/ai:ai_strategy",
+        "//src/main/cpp/net/eagle0/shardok/ai:ai_unit_score_calculator",
+        "//src/main/cpp/net/eagle0/shardok/ai:ai_water_crossing_calculator",
+        "//src/main/cpp/net/eagle0/shardok/ai/score/private:abstract_ai_score_calculator",
+        "//src/main/cpp/net/eagle0/shardok/ai/score/private:ai_score_calculator_shared_utilities",
+        "//src/main/cpp/net/eagle0/shardok/library:engine",
+        "//src/main/cpp/net/eagle0/shardok/library:game_state_w",
+        "//src/main/cpp/net/eagle0/shardok/library/map:coords_set",
+        "//src/main/cpp/net/eagle0/shardok/library/settings:game_settings",
+    ],
+)
+
 cc_library(
    name = "standard_ai_score_calculator",
    srcs = ["StandardAIScoreCalculator.cpp"],
@@ -54,7 +79,6 @@ cc_library(
    deps = [
        ":ai_score_calculator_interface",
        ":ai_victory_condition_score_calculator",
-        "//src/main/cpp/net/eagle0/shardok/ai:ai_experiment_config",
        "//src/main/cpp/net/eagle0/shardok/ai:ai_strategy",
        "//src/main/cpp/net/eagle0/shardok/ai:ai_unit_score_calculator",
        "//src/main/cpp/net/eagle0/shardok/ai:ai_water_crossing_calculator",
@@ -59,7 +59,6 @@ public:
    MCTSOptimizedAIScoreCalculator(
            int maxRounds,
            ActionPoints braveWaterCost,
-            ActionPoints archeryActionPointCost,
            int meteorRange,
            double meteorCastVigorCost,
            int minimumFleeOddsThreshold,
@@ -70,7 +69,6 @@ public:
        : AbstractAIScoreCalculator(
                  maxRounds,
                  braveWaterCost,
-                  archeryActionPointCost,
                  meteorRange,
                  meteorCastVigorCost,
                  minimumFleeOddsThreshold,
@@ -238,7 +236,6 @@ auto MakeMCTSOptimizedAIScoreCalculator(
    return std::make_unique<MCTSOptimizedAIScoreCalculator>(
            settingsGetter.Backing().max_rounds(),
            settingsGetter.Backing().brave_water_action_point_cost(),
-            settingsGetter.Backing().archery_action_point_cost(),
            settingsGetter.Backing().meteor_range(),
            settingsGetter.Backing().meteor_cast_vigor_cost(),
            settingsGetter.Backing().ai_minimum_flee_odds_threshold(),
@@ -5,7 +5,7 @@
 We need a scoring algorithm that makes MCTS perform well by providing score differences in the right range:

 - **Standard Scorer**: Returns unbounded relative scores. Small differences get amplified in MCTS UCB formula, causing over-exploitation (commits to 1-2 high-scoring nodes too early).
- **Retired bounded-ratio scorer**: Returned scores in [0,1] range with power transformation (exponent=0.1). Differences were too compressed (~0.02-0.05), causing over-exploration (all nodes explored equally, AI made bad choices).
+- **Normalized Scorer**: Returns scores in [0,1] range with power transformation (exponent=0.1). Differences are too compressed (~0.02-0.05), causing over-exploration (all nodes explored equally, AI makes bad choices).

 ### MCTS Requirements

@@ -32,9 +32,9 @@ This means individual scores should differ by **0.5 to 2.0** between meaningfull
 - Range: 0 (all captured) to -3000 (all held, far away)

 ### Terminal States
- Victory: INT_MAX
- Defeat: INT_MIN
- Flee/Draw: 0
+- Victory: INT_MAX (or 1.0 for normalized)
+- Defeat: INT_MIN (or 0.0 for normalized)
+- Flee/Draw: 0 (or 0.5 for normalized)
 - Captured unit: -10,000
 - Captured VIP: -25,000

@@ -257,7 +257,7 @@ This progression is ideal:

 This avoids both pathologies:
 - Not over-exploiting (like Standard scorer which overcommitted to tiny early differences)
- Not over-exploring like the retired bounded-ratio scorer, which explored equally even with large advantages
+- Not over-exploring (like Normalized scorer which explored equally even with large advantages)

 ## Why MCTS Needs Stronger Signal Than Minimax

@@ -281,11 +281,11 @@ This gives MCTS the right balance: explore when moves are similar, exploit when

 ## Implementation Notes

-1. **Use same calculation structure**: Inherit from AbstractAIScoreCalculator like Standard
+1. **Use same calculation structure**: Inherit from AbstractAIScoreCalculator like Standard and Normalized
 2. **Reuse unit scoring**: Use existing CalculateUnitsScoreComponents and victory condition calculators
 3. **Only change combination**: Override CombineAttackerScores, CombineDefenderScores, etc.
 4. **Bounded terminals**: Use ±1000 instead of INT_MAX/MIN for numerical stability
-5. **No transformation**: Do not apply power transformation - linear scaling is sufficient
+5. **No transformation**: Unlike Normalized, don't apply power transformation - linear scaling is sufficient
 6. **Scale constants tuned for MCTS**: 10x larger than naive normalization to provide appropriate signal strength

 ## Testing with Integration Tests
@@ -383,5 +383,6 @@ Once the scorer passes integration tests with IterativeDeepeningAI, then integra

 - `BoundedLinearAIScoreCalculator`
 - `MCTSOptimizedAIScoreCalculator`
+- `LinearNormalizedAIScoreCalculator`

 Recommend: **`MCTSOptimizedAIScoreCalculator`** to clearly indicate purpose.
@@ -0,0 +1,252 @@
+//
+// Normalized [0,1] implementation of AIScoreCalculator
+//
+
+#include "NormalizedAIScoreCalculator.hpp"
+
+#include <cmath>
+#include <unordered_map>
+
+#include "private/AIScoreCalculatorSharedUtilities.hpp"
+#include "private/AbstractAIScoreCalculator.hpp"
+#include "src/main/cpp/net/eagle0/shardok/ai/AIAttackLocations.hpp"
+#include "src/main/cpp/net/eagle0/shardok/ai/AIScoreUtilities.hpp"
+#include "src/main/cpp/net/eagle0/shardok/ai/AIStrategy.hpp"
+#include "src/main/cpp/net/eagle0/shardok/ai/AIUnitScoreCalculator.hpp"
+#include "src/main/cpp/net/eagle0/shardok/ai/score/AIScoreCalculator.hpp"
+#include "src/main/cpp/net/eagle0/shardok/library/ShardokEngine.hpp"
+#include "src/main/cpp/net/eagle0/shardok/library/action_point_distances/ActionPointDistancesCache.hpp"
+#include "src/main/cpp/net/eagle0/shardok/library/settings/GameSettings.hpp"
+
+namespace shardok {
+
+using net::eagle0::shardok::storage::fb::BattalionTypeId;
+
+// Bring shared utilities into scope
+using score_calculator_internal::UnitsScoreComponents;
+
+/// Normalized implementation of AIScoreCalculator that produces scores in [0, 1] range.
+/// Inherits from AbstractAIScoreCalculator to share common functionality.
+class NormalizedAIScoreCalculator : public AbstractAIScoreCalculator {
+public:
+    NormalizedAIScoreCalculator(
+            int maxRounds,
+            ActionPoints braveWaterCost,
+            int meteorRange,
+            double meteorCastVigorCost,
+            int minimumFleeOddsThreshold,
+            int desperateFleeThreshold,
+            std::vector<BattalionTypeSPtr> battalionTypes,
+            const APDCache &apdCache,
+            const ALCache &alCache)
+        : AbstractAIScoreCalculator(
+                  maxRounds,
+                  braveWaterCost,
+                  meteorRange,
+                  meteorCastVigorCost,
+                  minimumFleeOddsThreshold,
+                  desperateFleeThreshold,
+                  std::move(battalionTypes),
+                  apdCache,
+                  alCache) {}
+
+    [[nodiscard]] auto GuessedStateScore(
+            bool isDefender,
+            const GameStateW &state,
+            const AIStrategy &aiStrategy,
+            const CoordsSet &allCastleCoords) const -> ScoreValue override;
+
+    // Implement pure virtual methods from AbstractAIScoreCalculator
+    [[nodiscard]] auto InterpretDefenderOutcome(GameOutcome outcome) const -> ScoreValue override;
+    [[nodiscard]] auto InterpretAttackerOutcome(GameOutcome outcome) const -> ScoreValue override;
+
+    [[nodiscard]] auto AttackerFleeStrategyScoreForState(const GameStateW &gameState) const
+            -> ScoreValue override;
+
+    [[nodiscard]] auto CombineAttackerScores(
+            const UnitsScoreComponents &components,
+            double victoryConditionScore,
+            int roundsRemaining) const -> ScoreValue override;
+
+    [[nodiscard]] auto CombineDefenderScatterScores(const UnitsScoreComponents &components) const
+            -> ScoreValue override;
+
+    [[nodiscard]] auto CombineDefenderHoldCastlesScores(
+            const UnitsScoreComponents &components,
+            double victoryConditionScore,
+            int roundsRemaining) const -> ScoreValue override;
+
+private:
+    [[nodiscard]] auto DefenderFleeStrategyScoreForState(const GameStateW &gameState) const
+            -> ScoreValue override;
+
+    /// Applies power transformation to spread out compressed scores for MCTS.
+    /// Maps [0,1] → [0,1] but pushes values away from 0.5 toward the extremes.
+    /// Terminal states (0.0, 1.0) are unchanged.
+    [[nodiscard]] auto TransformForMCTS(ScoreValue score) const -> ScoreValue;
+};
+
+// Implementation of NormalizedAIScoreCalculator methods
+
+auto NormalizedAIScoreCalculator::InterpretDefenderOutcome(GameOutcome outcome) const
+        -> ScoreValue {
+    switch (outcome) {
+        case GameOutcome::DEFENDER_VICTORY: return 1.0;
+        case GameOutcome::ATTACKER_VICTORY: return 0.0;
+        case GameOutcome::DRAW: return 0.5;
+        case GameOutcome::FLEE_OUTCOME: return 0.5;
+    }
+    throw ShardokInternalErrorException("Unknown GameOutcome");
+}
+
+auto NormalizedAIScoreCalculator::InterpretAttackerOutcome(GameOutcome outcome) const
+        -> ScoreValue {
+    switch (outcome) {
+        case GameOutcome::ATTACKER_VICTORY: return 1.0;
+        case GameOutcome::DEFENDER_VICTORY: return 0.0;
+        case GameOutcome::DRAW: return 0.5;
+        case GameOutcome::FLEE_OUTCOME: return 0.5;
+    }
+    throw ShardokInternalErrorException("Unknown GameOutcome");
+}
+
+auto NormalizedAIScoreCalculator::CombineDefenderScatterScores(
+        const UnitsScoreComponents &components) const -> ScoreValue {
+    // For defender, we flip the perspective: defenderValue is (1), attackerValue is (2)
+    const double defenderValue = components.defenderUnitsValue;
+    const double attackerValue = components.attackerUnitsValue;
+
+    // No victory condition for scatter strategy
+    const double denominator = defenderValue + attackerValue;
+    if (denominator == 0.0) { return 0.5; }
+
+    return defenderValue / denominator;
+}
+
+auto NormalizedAIScoreCalculator::CombineDefenderHoldCastlesScores(
+        const UnitsScoreComponents &components,
+        const double victoryConditionScore,
+        const int /*roundsRemaining*/) const -> ScoreValue {
+    // For defender, flip perspective
+    const double defenderValue = components.defenderUnitsValue;
+    const double attackerValue = components.attackerUnitsValue;
+
+    // Apply normalization
+    double numerator;
+    double denominator;
+
+    if (victoryConditionScore >= 0) {
+        numerator = defenderValue + victoryConditionScore;
+        denominator = defenderValue + attackerValue + victoryConditionScore;
+    } else {
+        numerator = defenderValue;
+        denominator = defenderValue + attackerValue - victoryConditionScore;
+    }
+
+    if (denominator == 0.0) { return 0.5; }
+
+    return numerator / denominator;
+}
+
+auto NormalizedAIScoreCalculator::DefenderFleeStrategyScoreForState(
+        const GameStateW & /*gameState*/) const -> ScoreValue {
+    // FLEE strategy doesn't fit the [0,1] model well - return 0.5
+    return 0.5;
+}
+
+auto NormalizedAIScoreCalculator::AttackerFleeStrategyScoreForState(
+        const GameStateW & /*gameState*/) const -> ScoreValue {
+    // FLEE strategy doesn't fit the [0,1] model well - return 0.5
+    return 0.5;
+}
+
+auto NormalizedAIScoreCalculator::CombineAttackerScores(
+        const UnitsScoreComponents &components,
+        const double victoryConditionScore,
+        const int /*roundsRemaining*/) const -> ScoreValue {
+    const double attackerUnitsValue = components.attackerUnitsValue;
+    const double defenderUnitsValue = components.defenderUnitsValue;
+
+    // Apply normalization formula
+    double numerator;
+    double denominator;
+
+    if (victoryConditionScore >= 0) {
+        // Positive victory condition: add to numerator
+        numerator = attackerUnitsValue + victoryConditionScore;
+        denominator = attackerUnitsValue + defenderUnitsValue + victoryConditionScore;
+    } else {
+        // Negative victory condition: subtract from denominator (making it larger)
+        numerator = attackerUnitsValue;
+        denominator = attackerUnitsValue + defenderUnitsValue - victoryConditionScore;
+    }
+
+    // Handle edge case of all zeros
+    if (denominator == 0.0) { return 0.5; }
+
+    return numerator / denominator;
+}
+
+auto NormalizedAIScoreCalculator::TransformForMCTS(ScoreValue score) const -> ScoreValue {
+    // Power transformation exponent - lower values spread scores more toward extremes
+    // Tuned for MCTS: balances exploration vs exploitation
+    // - Too low (e.g., 0.3): over-exploitation like standard scorer
+    // - Too high (e.g., 0.9): over-exploration like untransformed normalized
+    // - 0.6-0.7: sweet spot for MCTS
+    constexpr double EXPONENT = 0.1;
+
+    if (score > 0.5) {
+        // Map [0.5, 1.0] → [0.5, 1.0] with power curve
+        // (score - 0.5) * 2.0 maps to [0, 1], apply power, then scale back
+        return 0.5 + 0.5 * std::pow((score - 0.5) * 2.0, EXPONENT);
+    } else {
+        // Map [0.0, 0.5] → [0.0, 0.5] with power curve (symmetric)
+        return 0.5 - 0.5 * std::pow((0.5 - score) * 2.0, EXPONENT);
+    }
+}
+
+auto NormalizedAIScoreCalculator::GuessedStateScore(
+        const bool isDefender,
+        const GameStateW &state,
+        const AIStrategy &aiStrategy,
+        const CoordsSet &allCastleCoords) const -> ScoreValue {
+    const int roundsRemaining = GetMaxRounds() - state->current_round();
+
+    ScoreValue rawScore;
+    if (isDefender) {
+        rawScore = DefenderScoreForState(state, aiStrategy, allCastleCoords, roundsRemaining);
+    } else {
+        rawScore = AttackerScoreForState(state, aiStrategy, allCastleCoords, roundsRemaining);
+    }
+
+    // Apply power transformation to spread out scores for MCTS
+    return TransformForMCTS(rawScore);
+}
+
+// Factory function implementation
+auto MakeNormalizedAIScoreCalculator(
+        const SettingsGetter &settingsGetter,
+        const APDCache &apdCache,
+        const ALCache &alCache) -> std::unique_ptr<AIScoreCalculator> {
+    // Extract all battalion types into a vector indexed by BattalionTypeId
+    std::vector<BattalionTypeSPtr> battalionTypes(BattalionTypeId::BattalionTypeId_MAX + 1);
+    for (int typeId = BattalionTypeId::BattalionTypeId_MIN;
+         typeId <= BattalionTypeId::BattalionTypeId_MAX;
+         typeId++) {
+        auto battalionTypeId = static_cast<BattalionTypeId>(typeId);
+        battalionTypes[battalionTypeId] = settingsGetter.GetBattalionType(battalionTypeId);
+    }
+
+    return std::make_unique<NormalizedAIScoreCalculator>(
+            settingsGetter.Backing().max_rounds(),
+            settingsGetter.Backing().brave_water_action_point_cost(),
+            settingsGetter.Backing().meteor_range(),
+            settingsGetter.Backing().meteor_cast_vigor_cost(),
+            settingsGetter.Backing().ai_minimum_flee_odds_threshold(),
+            settingsGetter.Backing().ai_desperate_flee_threshold(),
+            std::move(battalionTypes),
+            apdCache,
+            alCache);
+}
+
+}  // namespace shardok
@@ -0,0 +1,44 @@
+//
+// Normalized [0,1] implementation of AIScoreCalculator
+//
+
+#ifndef EAGLE0_NORMALIZEDAISCORECALCULATOR_HPP
+#define EAGLE0_NORMALIZEDAISCORECALCULATOR_HPP
+
+#include <memory>
+
+#include "src/main/cpp/net/eagle0/shardok/ai/AIAttackLocations.hpp"
+#include "src/main/cpp/net/eagle0/shardok/library/action_point_distances/ActionPointDistancesCache.hpp"
+#include "src/main/cpp/net/eagle0/shardok/library/settings/GameSettings.hpp"
+
+namespace shardok {
+
+// Forward declarations
+class AIScoreCalculator;
+
+using APDCache = std::shared_ptr<ActionPointDistancesCache>;
+using ALCache = std::unique_ptr<AttackLocationsCache>;
+
+/// Factory function to create a NormalizedAIScoreCalculator.
+/// Returns a unique_ptr to AIScoreCalculator to hide the implementation.
+///
+/// The normalized scorer produces scores in the range [0, 1] where:
+/// - 0.0 = complete defender victory
+/// - 1.0 = complete attacker victory
+/// - 0.5 = neutral/draw state
+///
+/// Terminal states (victory/defeat) always return 1.0 or 0.0.
+/// Non-terminal states use asymmetric normalization:
+/// - If victory condition >= 0:
+///   score = (attackerUnits + victoryCondition) / (attackerUnits + defenderUnits +
+///   victoryCondition)
+/// - If victory condition < 0:
+///   score = attackerUnits / (attackerUnits + defenderUnits - victoryCondition)
+[[nodiscard]] auto MakeNormalizedAIScoreCalculator(
+        const SettingsGetter& settingsGetter,
+        const APDCache& apdCache,
+        const ALCache& alCache) -> std::unique_ptr<AIScoreCalculator>;
+
+}  // namespace shardok
+
+#endif  // EAGLE0_NORMALIZEDAISCORECALCULATOR_HPP
@@ -4,17 +4,14 @@

 #include "StandardAIScoreCalculator.hpp"

-#include <algorithm>
 #include <atomic>
 #include <chrono>
-#include <cmath>
 #include <unordered_map>

 #include "private/AIScoreCalculatorSharedUtilities.hpp"
 #include "private/AbstractAIScoreCalculator.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai/AIAttackGroups.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai/AIAttackLocations.hpp"
-#include "src/main/cpp/net/eagle0/shardok/ai/AIExperimentConfig.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai/AIScoreUtilities.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai/AIStrategy.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai/AIUnitScoreCalculator.hpp"
@@ -70,68 +67,6 @@ struct AttackerScorePerformanceLogger {
    }
 };

-struct StandardScoringExperiment {
-    double attackerUnitsScale = 1.0;
-    double defenderUnitsScale = 1.0;
-    double victoryConditionScale = 1.0;
-    double unitsBaseMultiplierScale = 1.0;
-    double attackerRoundMultiplier = 1.0;
-    double defenderRoundAdvantageScale = 1.0;
-};
-
-auto ExperimentalScoringId() -> int { return CurrentAIExperimentId(); }
-
-auto ScoringExperimentForId(const int id) -> StandardScoringExperiment {
-    StandardScoringExperiment experiment;
-
-    switch (id) {
-        case 1: experiment.attackerUnitsScale = 1.05; break;
-        case 2: experiment.attackerUnitsScale = 1.10; break;
-        case 3: experiment.attackerUnitsScale = 0.95; break;
-        case 4: experiment.attackerUnitsScale = 0.90; break;
-        case 5: experiment.defenderUnitsScale = 0.95; break;
-        case 6: experiment.defenderUnitsScale = 0.90; break;
-        case 7: experiment.defenderUnitsScale = 1.05; break;
-        case 8: experiment.defenderUnitsScale = 1.10; break;
-        case 9: experiment.victoryConditionScale = 1.10; break;
-        case 10: experiment.victoryConditionScale = 1.25; break;
-        case 11: experiment.victoryConditionScale = 0.90; break;
-        case 12: experiment.victoryConditionScale = 0.75; break;
-        case 13: experiment.unitsBaseMultiplierScale = 0.90; break;
-        case 14: experiment.unitsBaseMultiplierScale = 1.10; break;
-        case 15: experiment.attackerRoundMultiplier = 0.90; break;
-        case 16: experiment.attackerRoundMultiplier = 0.75; break;
-        case 17: experiment.defenderRoundAdvantageScale = 0.50; break;
-        case 18: experiment.defenderRoundAdvantageScale = 0.0; break;
-        case 19:
-            experiment.attackerUnitsScale = 1.10;
-            experiment.victoryConditionScale = 1.10;
-            break;
-        case 20:
-            experiment.defenderUnitsScale = 0.90;
-            experiment.victoryConditionScale = 1.10;
-            break;
-        default: break;
-    }
-
-    return experiment;
-}
-
-auto AdjustDefenderUnitsForExperiment(
-        double defenderUnitsValue,
-        const int roundsRemaining,
-        const int maxRounds,
-        const StandardScoringExperiment &experiment) -> double {
-    const double currentRound = static_cast<double>(maxRounds - roundsRemaining);
-    const double standardAdvantage = 1.0 + currentRound / 31.0;
-    if (standardAdvantage <= 0.0) { return defenderUnitsValue * experiment.defenderUnitsScale; }
-
-    const double unadjustedDefenderValue = defenderUnitsValue / standardAdvantage;
-    const double adjustedAdvantage =
-            1.0 + experiment.defenderRoundAdvantageScale * currentRound / 31.0;
-    return unadjustedDefenderValue * adjustedAdvantage * experiment.defenderUnitsScale;
-}
-
 std::atomic<int> AttackerScorePerformanceLogger::callCount{0};
 std::atomic<double> AttackerScorePerformanceLogger::intervalTime{0.0};
 std::atomic<double> AttackerScorePerformanceLogger::totalTime{0.0};
@@ -160,27 +95,23 @@ public:
    StandardAIScoreCalculator(
            int maxRounds,
            ActionPoints braveWaterCost,
-            ActionPoints archeryActionPointCost,
            int meteorRange,
            double meteorCastVigorCost,
            int minimumFleeOddsThreshold,
            int desperateFleeThreshold,
            std::vector<BattalionTypeSPtr> battalionTypes,
            const APDCache &apdCache,
-            const ALCache &alCache,
-            StandardScoringExperiment experiment = {})
+            const ALCache &alCache)
        : AbstractAIScoreCalculator(
                  maxRounds,
                  braveWaterCost,
-                  archeryActionPointCost,
                  meteorRange,
                  meteorCastVigorCost,
                  minimumFleeOddsThreshold,
                  desperateFleeThreshold,
                  std::move(battalionTypes),
                  apdCache,
-                  alCache),
-          experiment_(experiment) {}
+                  alCache) {}

    [[nodiscard]] auto GuessedStateScore(
            bool isDefender,
@@ -220,8 +151,6 @@ private:

    [[nodiscard]] auto DefenderFleeStrategyScoreForState(const GameStateW &gameState) const
            -> ScoreValue override;
-
-    StandardScoringExperiment experiment_;
 };

 // Implementation of StandardAIScoreCalculator methods
@@ -230,7 +159,7 @@ auto StandardAIScoreCalculator::InterpretDefenderOutcome(GameOutcome outcome) co
    switch (outcome) {
        case GameOutcome::DEFENDER_VICTORY: return INT_MAX;
        case GameOutcome::ATTACKER_VICTORY: return INT_MIN;
-        case GameOutcome::DRAW: return INT_MAX;
+        case GameOutcome::DRAW: return 0;
        case GameOutcome::FLEE_OUTCOME: return 0;
    }
    throw ShardokInternalErrorException("Unknown GameOutcome");
@@ -240,7 +169,7 @@ auto StandardAIScoreCalculator::InterpretAttackerOutcome(GameOutcome outcome) co
    switch (outcome) {
        case GameOutcome::ATTACKER_VICTORY: return INT_MAX;
        case GameOutcome::DEFENDER_VICTORY: return INT_MIN;
-        case GameOutcome::DRAW: return INT_MIN;
+        case GameOutcome::DRAW: return 0;
        case GameOutcome::FLEE_OUTCOME: return 0;
    }
    throw ShardokInternalErrorException("Unknown GameOutcome");
@@ -312,20 +241,10 @@ auto StandardAIScoreCalculator::CombineAttackerScores(
        const UnitsScoreComponents &components,
        const double victoryConditionScore,
        const int roundsRemaining) const -> ScoreValue {
-    const double attackerUnitsValue =
-            components.attackerUnitsValue * experiment_.attackerUnitsScale;
-    const double defenderUnitsValue = AdjustDefenderUnitsForExperiment(
-            components.defenderUnitsValue,
-            roundsRemaining,
-            GetMaxRounds(),
-            experiment_);
-    const double unitsDifference = attackerUnitsValue - defenderUnitsValue;
-    // Keep tactical unit value relevant late in the battle. The previous time decay could make
-    // attackers undervalue decisive combat near the round limit.
-    const double unitsMultiplier = experiment_.attackerRoundMultiplier;
-    return UNITS_BASE_MULTIPLIER * experiment_.unitsBaseMultiplierScale * unitsMultiplier *
-                   unitsDifference +
-           victoryConditionScore * experiment_.victoryConditionScale;
+    const double unitsDifference = components.attackerUnitsValue - components.defenderUnitsValue;
+    const double unitsMultiplier =
+            static_cast<double>(roundsRemaining) / static_cast<double>(GetMaxRounds());
+    return UNITS_BASE_MULTIPLIER * unitsMultiplier * unitsDifference + victoryConditionScore;
 }

 auto StandardAIScoreCalculator::GuessedStateScore(
@@ -342,11 +261,10 @@ auto StandardAIScoreCalculator::GuessedStateScore(
 }

 // Factory function implementation
-auto MakeStandardAIScoreCalculatorWithExperiment(
+auto MakeStandardAIScoreCalculator(
        const SettingsGetter &settingsGetter,
        const APDCache &apdCache,
-        const ALCache &alCache,
-        const StandardScoringExperiment &experiment) -> std::unique_ptr<AIScoreCalculator> {
+        const ALCache &alCache) -> std::unique_ptr<AIScoreCalculator> {
    // Extract all battalion types into a vector indexed by BattalionTypeId
    std::vector<BattalionTypeSPtr> battalionTypes(BattalionTypeId::BattalionTypeId_MAX + 1);
    for (int typeId = BattalionTypeId::BattalionTypeId_MIN;
@@ -359,37 +277,13 @@ auto MakeStandardAIScoreCalculatorWithExperiment(
    return std::make_unique<StandardAIScoreCalculator>(
            settingsGetter.Backing().max_rounds(),
            settingsGetter.Backing().brave_water_action_point_cost(),
-            settingsGetter.Backing().archery_action_point_cost(),
            settingsGetter.Backing().meteor_range(),
            settingsGetter.Backing().meteor_cast_vigor_cost(),
            settingsGetter.Backing().ai_minimum_flee_odds_threshold(),
            settingsGetter.Backing().ai_desperate_flee_threshold(),
            std::move(battalionTypes),
            apdCache,
-            alCache,
-            experiment);
-}
-
-auto MakeStandardAIScoreCalculator(
-        const SettingsGetter &settingsGetter,
-        const APDCache &apdCache,
-        const ALCache &alCache) -> std::unique_ptr<AIScoreCalculator> {
-    return MakeStandardAIScoreCalculatorWithExperiment(
-            settingsGetter,
-            apdCache,
-            alCache,
-            StandardScoringExperiment{});
-}
-
-auto MakeExperimentalAIScoreCalculator(
-        const SettingsGetter &settingsGetter,
-        const APDCache &apdCache,
-        const ALCache &alCache) -> std::unique_ptr<AIScoreCalculator> {
-    return MakeStandardAIScoreCalculatorWithExperiment(
-            settingsGetter,
-            apdCache,
-            alCache,
-            ScoringExperimentForId(ExperimentalScoringId()));
+            alCache);
 }

 }  // namespace shardok
@@ -26,12 +26,6 @@ using ALCache = std::unique_ptr<AttackLocationsCache>;
        const APDCache& apdCache,
        const ALCache& alCache) -> std::unique_ptr<AIScoreCalculator>;

-/// Factory function to create the environment-selected experimental scorer.
-[[nodiscard]] auto MakeExperimentalAIScoreCalculator(
-        const SettingsGetter& settingsGetter,
-        const APDCache& apdCache,
-        const ALCache& alCache) -> std::unique_ptr<AIScoreCalculator>;
-
 }  // namespace shardok

 #endif  // EAGLE0_STANDARDAISCORECALCULATOR_HPP
@@ -40,7 +40,7 @@ constexpr double kMaxProximityBuf = 1.5;
 constexpr double kDistanceDebufRatio = 8.0;

 /// Structure to hold separated attacker/defender unit scores
-/// Used by scorers that combine attacker and defender unit values separately.
+/// Used by NormalizedAIScoreCalculator to apply asymmetric normalization
 struct UnitsScoreComponents {
    double attackerUnitsValue;
    double defenderUnitsValue;
@@ -157,8 +157,7 @@ auto AbstractAIScoreCalculator::CalculateUnitsScoreComponents(
                locationsCausingDanger,
                notBravingApd,
                GetMeteorRange(),
-                GetMeteorCastVigorCost(),
-                GetArcheryActionPointCost());
+                GetMeteorCastVigorCost());

        attackerUnitsValue += distanceMultiplier * uv;
    }
@@ -189,8 +188,7 @@ auto AbstractAIScoreCalculator::CalculateUnitsScoreComponents(
                locationsCausingDangerForAttacker,
                defenderNotBravingApd,
                GetMeteorRange(),
-                GetMeteorCastVigorCost(),
-                GetArcheryActionPointCost());
+                GetMeteorCastVigorCost());

        double distanceMultiplier = 1.0;

@@ -253,7 +251,7 @@ auto AbstractAIScoreCalculator::CalculateUnitsScoreComponents(
                                                      braveWaterCost)
                                            : nullptr,
                                    cachedHexMap);
-                            if (thisDistance < closestDistanceToFriendly) {
+                            if (thisDistance < closestDistanceToEnemy) {
                                closestDistanceToFriendly = thisDistance;
                            }
                        }
@@ -508,7 +506,9 @@ auto AbstractAIScoreCalculator::AttackerScoreForState(
    const ScoreValue victoryConditionTotal =
            CalculateAttackerVictoryConditionScore(gameState, attackerStrategy, castleCoords);

-    // Combine scores using subclass-specific logic.
+    // Combine scores using subclass-specific logic
+    // Standard: uses difference and multiplier
+    // Normalized: uses normalization formula
    return CombineAttackerScores(components, victoryConditionTotal, roundsRemaining);
 }

@@ -44,7 +44,6 @@ protected:
    AbstractAIScoreCalculator(
            int maxRounds,
            ActionPoints braveWaterCost,
-            ActionPoints archeryActionPointCost,
            int meteorRange,
            double meteorCastVigorCost,
            int minimumFleeOddsThreshold,
@@ -54,7 +53,6 @@ protected:
            const ALCache &alCache)
        : maxRounds_(maxRounds),
          braveWaterCost_(braveWaterCost),
-          archeryActionPointCost_(archeryActionPointCost),
          meteorRange_(meteorRange),
          meteorCastVigorCost_(meteorCastVigorCost),
          minimumFleeOddsThreshold_(minimumFleeOddsThreshold),
@@ -72,9 +70,6 @@ protected:
    [[nodiscard]] auto GetApdCache() const -> const APDCache & { return apdCache_; }
    [[nodiscard]] auto GetAlCache() const -> const ALCache & { return alCache_; }
    [[nodiscard]] auto GetBraveWaterCost() const -> ActionPoints { return braveWaterCost_; }
-    [[nodiscard]] auto GetArcheryActionPointCost() const -> ActionPoints {
-        return archeryActionPointCost_;
-    }
    [[nodiscard]] auto GetMaxRounds() const -> int { return maxRounds_; }
    [[nodiscard]] auto GetMeteorRange() const -> int { return meteorRange_; }
    [[nodiscard]] auto GetMeteorCastVigorCost() const -> double { return meteorCastVigorCost_; }
@@ -151,8 +146,9 @@ protected:
            -> ScoreValue = 0;

    // Pure virtual method for combining units score and victory condition score
-    // This is where concrete scorers combine unit value and victory condition pressure.
-    // Standard uses the raw unit-value difference; MCTS-optimized uses bounded scaling.
+    // This is where Standard and Normalized diverge in their scoring approach
+    // Standard: uses difference and applies multiplier
+    // Normalized: uses normalization formula with separate attacker/defender values
    [[nodiscard]] virtual auto CombineAttackerScores(
            const UnitsScoreComponents &components,
            double victoryConditionScore,
@@ -162,7 +158,6 @@ private:
    // Scalar settings extracted from SettingsGetter
    int maxRounds_;
    ActionPoints braveWaterCost_;
-    ActionPoints archeryActionPointCost_;
    int meteorRange_;
    double meteorCastVigorCost_;
    int minimumFleeOddsThreshold_;
@@ -54,38 +54,30 @@ std::unique_ptr<BattleConfigProto> AiBattleConfigLoader::CreateDefaultPerfConfig
    // Basic setup matching Unity's "Perf" configuration
    config->set_map_name("Alah");
    config->set_month(month);
-    config->set_max_rounds(31);
-    config->set_random_seed(1);
-    config->set_use_configured_placements(false);
-    config->set_apply_forced_commands(true);
+    config->set_max_rounds(40);
+    config->set_random_seed(0);  // Use system random

    // Attacker configuration - 6 Longbowmen units with professions 1-6
    PlayerConfig* attacker = config->mutable_attacker();
    attacker->set_ai_algorithm(AIAlgorithmType::ITERATIVE_DEEPENING);
-    attacker->set_scoring_calculator(
-            net::eagle0::shardok::ai_battle_simulator::ScoringCalculatorType::STANDARD);

    for (int profession = 1; profession <= 6; ++profession) {
        UnitConfig* unit = attacker->add_units();
        unit->set_profession(profession);
-        unit->set_battalion_type_id(4);        // Longbowmen (battalion type 4)
-        unit->set_starting_position_index(0);  // Attackers use position 0
-        unit->set_can_flee(true);
+        unit->set_battalion_type_id(4);             // Longbowmen (battalion type 4)
+        unit->set_starting_position_index(0);       // Attackers use position 0
        unit->mutable_battalion()->set_size(1000);  // Match client battalion size
    }

    // Defender configuration - 6 Light Infantry units with NO_PROFESSION (profession 14)
    PlayerConfig* defender = config->mutable_defender();
    defender->set_ai_algorithm(AIAlgorithmType::ITERATIVE_DEEPENING);
-    defender->set_scoring_calculator(
-            net::eagle0::shardok::ai_battle_simulator::ScoringCalculatorType::STANDARD);

    for (int i = 0; i < 6; ++i) {
        UnitConfig* unit = defender->add_units();
-        unit->set_profession(14);               // NO_PROFESSION to match client battles
-        unit->set_battalion_type_id(0);         // Light Infantry (battalion type 0)
-        unit->set_starting_position_index(-1);  // Defenders use position -1
-        unit->set_can_flee(true);
+        unit->set_profession(14);                   // NO_PROFESSION to match client battles
+        unit->set_battalion_type_id(0);             // Light Infantry (battalion type 0)
+        unit->set_starting_position_index(-1);      // Defenders use position -1
        unit->mutable_battalion()->set_size(1000);  // Match client battalion size
    }

@@ -4,19 +4,14 @@

 #include "AiBattleSimulator.hpp"

-#include <algorithm>
 #include <iostream>
-#include <sstream>
-#include <stdexcept>

 #include "src/main/cpp/net/eagle0/common/FilesystemUtils.hpp"
-#include "src/main/cpp/net/eagle0/common/RandomGenerator.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai/ShardokAIClient.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai_testing_common/AIClientFactory.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai_testing_common/GamePhaseRunner.hpp"
 #include "src/main/cpp/net/eagle0/shardok/ai_testing_common/GameSettingsFactory.hpp"
 #include "src/main/cpp/net/eagle0/shardok/library/ShardokEngine.hpp"
-#include "src/main/cpp/net/eagle0/shardok/library/UnitPlacementInfo.hpp"
 #include "src/main/cpp/net/eagle0/shardok/library/fb_helpers/GameStateHelpers.hpp"
 #include "src/main/cpp/net/eagle0/shardok/library/settings/GameSettings.hpp"
 #include "src/main/cpp/net/eagle0/shardok/util/MapLoader.hpp"
@@ -50,67 +45,6 @@ constexpr int DEFAULT_HERO_INTEGRITY = 0;
 constexpr int DEFAULT_HERO_AMBITION = 0;
 constexpr int DEFAULT_HERO_VIGOR = 102;

-constexpr int PROFESSION_MAGE = 1;
-constexpr double MINIMUM_AGILITY_FOR_ARCHERY = 75.0;
-constexpr double MINIMUM_AGILITY_FOR_START_FIRE = 70.0;
-constexpr double MINIMUM_WISDOM_FOR_START_FIRE = 70.0;
-
-struct SideTotals {
-    int attackerUnits = 0;
-    int defenderUnits = 0;
-    double attackerTroops = 0.0;
-    double defenderTroops = 0.0;
-};
-
-std::string JsonEscape(const std::string& value) {
-    std::ostringstream out;
-    for (const char c : value) {
-        switch (c) {
-            case '"': out << "\\\""; break;
-            case '\\': out << "\\\\"; break;
-            case '\n': out << "\\n"; break;
-            case '\r': out << "\\r"; break;
-            case '\t': out << "\\t"; break;
-            default: out << c; break;
-        }
-    }
-    return out.str();
-}
-
-std::string CompletionReasonLabel(EvaluationCompletionReason reason) {
-    switch (reason) {
-        case EvaluationCompletionReason::RAN_OUT_OF_COMMANDS: return "ran_out_of_commands";
-        case EvaluationCompletionReason::RAN_OUT_OF_TIME: return "ran_out_of_time";
-        case EvaluationCompletionReason::NOT_ENOUGH_TIME_TO_CONTINUE:
-            return "not_enough_time_to_continue";
-    }
-}
-
-SideTotals CalculateSideTotals(const GameStateW& state) {
-    SideTotals totals;
-    const auto* units = state->units();
-    if (units == nullptr) { return totals; }
-
-    for (const auto* unit : *units) {
-        if (unit == nullptr) { continue; }
-        const bool isAlive =
-                unit->status() == net::eagle0::shardok::storage::fb::UnitStatus_NORMAL_UNIT ||
-                unit->status() == net::eagle0::shardok::storage::fb::UnitStatus_RESERVE_UNIT ||
-                unit->status() == net::eagle0::shardok::storage::fb::UnitStatus_NEVER_ENTERED_UNIT;
-        if (!isAlive) { continue; }
-
-        if (unit->player_id() == ATTACKER_ID) {
-            totals.attackerUnits++;
-            totals.attackerTroops += unit->battalion().size();
-        } else if (unit->player_id() == DEFENDER_ID) {
-            totals.defenderUnits++;
-            totals.defenderTroops += unit->battalion().size();
-        }
-    }
-
-    return totals;
-}
-
 // Convert proto AI algorithm type to ShardokAI enum
 AIAlgorithmType ConvertAIAlgorithmType(
        net::eagle0::shardok::ai_battle_simulator::AIAlgorithmType protoType) {
@@ -121,160 +55,21 @@ AIAlgorithmType ConvertAIAlgorithmType(
    }
 }

-ScoringCalculatorType ConvertScoringCalculatorType(
-        net::eagle0::shardok::ai_battle_simulator::ScoringCalculatorType protoType) {
-    switch (protoType) {
-        case net::eagle0::shardok::ai_battle_simulator::EXPERIMENTAL:
-            return ScoringCalculatorType::EXPERIMENTAL;
-        case net::eagle0::shardok::ai_battle_simulator::MCTS_OPTIMIZED:
-            return ScoringCalculatorType::MCTS_OPTIMIZED;
-        case net::eagle0::shardok::ai_battle_simulator::STANDARD:
-        case net::eagle0::shardok::ai_battle_simulator::SCORING_CALCULATOR_TYPE_UNSPECIFIED:
-        default: return ScoringCalculatorType::STANDARD;
-    }
-}
-
-mcts::MCTSSimulationPolicy ConvertSimulationPolicy(
-        net::eagle0::shardok::ai_battle_simulator::MctsSimulationPolicy protoType) {
-    switch (protoType) {
-        case net::eagle0::shardok::ai_battle_simulator::RANDOM:
-            return mcts::MCTSSimulationPolicy::RANDOM;
-        case net::eagle0::shardok::ai_battle_simulator::FILTERED_RANDOM:
-            return mcts::MCTSSimulationPolicy::FILTERED_RANDOM;
-        case net::eagle0::shardok::ai_battle_simulator::BEST_IMMEDIATE:
-            return mcts::MCTSSimulationPolicy::BEST_IMMEDIATE;
-        case net::eagle0::shardok::ai_battle_simulator::WEIGHTED_BEST_IMMEDIATE:
-            return mcts::MCTSSimulationPolicy::WEIGHTED_BEST_IMMEDIATE;
-        case net::eagle0::shardok::ai_battle_simulator::WEIGHTED_HEURISTIC:
-        case net::eagle0::shardok::ai_battle_simulator::MCTS_SIMULATION_POLICY_UNSPECIFIED:
-        default: return mcts::MCTSSimulationPolicy::WEIGHTED_HEURISTIC;
-    }
-}
-
-mcts::MCTSBackpropagationPolicy ConvertBackpropagationPolicy(
-        net::eagle0::shardok::ai_battle_simulator::MctsBackpropagationPolicy protoType) {
-    switch (protoType) {
-        case net::eagle0::shardok::ai_battle_simulator::AVERAGING:
-            return mcts::MCTSBackpropagationPolicy::AVERAGING;
-        case net::eagle0::shardok::ai_battle_simulator::MINIMAX:
-        case net::eagle0::shardok::ai_battle_simulator::MCTS_BACKPROPAGATION_POLICY_UNSPECIFIED:
-        default: return mcts::MCTSBackpropagationPolicy::MINIMAX;
-    }
-}
-
-mcts::MCTSConfig BuildMCTSConfig(
-        const net::eagle0::shardok::ai_battle_simulator::PlayerConfig& playerConfig,
-        int battleSeed,
-        PlayerId playerId) {
-    mcts::MCTSConfig config{};
-    config.backpropagationPolicy = mcts::MCTSBackpropagationPolicy::MINIMAX;
-    config.maxPlayerFlips = 1;
-    config.maxSimulationFlips = 1;
-    config.simulationPolicy = mcts::MCTSSimulationPolicy::WEIGHTED_HEURISTIC;
-    if (battleSeed != 0) { config.useMultithreading = false; }
-    if (battleSeed != 0) {
-        config.randomSeed = static_cast<uint32_t>(battleSeed + (1009 * playerId));
-    }
-
-    if (!playerConfig.has_mcts_config()) { return config; }
-
-    const auto& protoConfig = playerConfig.mcts_config();
-    if (protoConfig.exploration_constant() > 0.0) {
-        config.explorationConstant = protoConfig.exploration_constant();
-    }
-    if (protoConfig.max_simulation_depth() > 0) {
-        config.maxSimulationDepth = protoConfig.max_simulation_depth();
-    }
-    if (protoConfig.max_tree_depth() > 0) { config.maxTreeDepth = protoConfig.max_tree_depth(); }
-    config.useMultithreading = protoConfig.use_multithreading();
-    if (protoConfig.num_threads() > 0) { config.numThreads = protoConfig.num_threads(); }
-    config.simulationPolicy = ConvertSimulationPolicy(protoConfig.simulation_policy());
-    config.backpropagationPolicy =
-            ConvertBackpropagationPolicy(protoConfig.backpropagation_policy());
-    if (protoConfig.max_player_flips() >= 0) {
-        config.maxPlayerFlips = protoConfig.max_player_flips();
-    }
-    if (protoConfig.max_simulation_flips() >= 0) {
-        config.maxSimulationFlips = protoConfig.max_simulation_flips();
-    }
-    if (protoConfig.random_seed() != 0) {
-        config.randomSeed = static_cast<uint32_t>(protoConfig.random_seed());
-    }
-    return config;
-}
-
-bool HasAttachedHero(int profession) { return profession > 0; }
-
 // Get value with default if not set (proto3 uses 0 as default, we check for that)
 template<typename T>
 T GetOrDefault(T value, T defaultValue) {
    return (value == 0) ? defaultValue : value;
 }

-bool HeroCanArchery(const net::eagle0::shardok::ai_battle_simulator::UnitConfig& unitConfig) {
-    if (!HasAttachedHero(unitConfig.profession())) { return false; }
-    return GetOrDefault(unitConfig.hero().agility(), DEFAULT_HERO_AGILITY) >=
-           MINIMUM_AGILITY_FOR_ARCHERY;
-}
-
-bool HeroCanStartFire(const net::eagle0::shardok::ai_battle_simulator::UnitConfig& unitConfig) {
-    if (!HasAttachedHero(unitConfig.profession())) { return false; }
-    return unitConfig.profession() == PROFESSION_MAGE ||
-           (GetOrDefault(unitConfig.hero().agility(), DEFAULT_HERO_AGILITY) >=
-                    MINIMUM_AGILITY_FOR_START_FIRE &&
-            GetOrDefault(unitConfig.hero().wisdom(), DEFAULT_HERO_WISDOM) >=
-                    MINIMUM_WISDOM_FOR_START_FIRE);
-}
-
-double GetBattalionSizeOrDefault(
-        const net::eagle0::shardok::ai_battle_simulator::BattalionConfig& battalion,
-        double defaultValue) {
-    return battalion.has_size() ? battalion.size() : defaultValue;
-}
-
-bool GetCanFleeOrDefault(const net::eagle0::shardok::ai_battle_simulator::UnitConfig& unitConfig) {
-    return unitConfig.has_can_flee() ? unitConfig.can_flee() : true;
-}
-
-double GetBattalionArmamentOrDefault(
-        const net::eagle0::shardok::ai_battle_simulator::BattalionConfig& battalion) {
-    return battalion.has_armament() ? battalion.armament() : DEFAULT_BATTALION_ARMAMENT;
-}
-
-double GetBattalionTrainingOrDefault(
-        const net::eagle0::shardok::ai_battle_simulator::BattalionConfig& battalion) {
-    return battalion.has_training() ? battalion.training() : DEFAULT_BATTALION_TRAINING;
-}
-
-double GetBattalionMoraleOrDefault(
-        const net::eagle0::shardok::ai_battle_simulator::BattalionConfig& battalion) {
-    return battalion.has_morale() ? battalion.morale() : DEFAULT_BATTALION_MORALE;
-}
-
 }  // namespace

-AiBattleSimulator::AiBattleSimulator(
-        const BattleConfigProto& config,
-        GameSettingsSPtr gameSettings,
-        std::optional<GameStateW> initialGameState,
-        BattleTraceOptions traceOptions)
+AiBattleSimulator::AiBattleSimulator(const BattleConfigProto& config, GameSettingsSPtr gameSettings)
    : config_(config),
-      gameSettings_(std::move(gameSettings)),
-      initialGameState_(std::move(initialGameState)),
-      traceOptions_(std::move(traceOptions)) {
+      gameSettings_(std::move(gameSettings)) {
    if (!gameSettings_) {
        // Initialize default game settings
        gameSettings_ = ai_testing_common::GameSettingsFactory::CreateDefault();
    }
-    if (config_.random_seed() != 0) {
-        auto setter = gameSettings_->GetSetter();
-        setter.SetRandomGenerator(std::make_shared<StdLibraryGenerator>(
-                static_cast<std::mt19937_64::result_type>(config_.random_seed())));
-    }
-    if (config_.max_rounds() > 0) {
-        auto setter = gameSettings_->GetSetter();
-        setter.SetInt("maxRounds", config_.max_rounds());
-    }
 }

 BattleResult AiBattleSimulator::RunBattle() {
@@ -282,7 +77,6 @@ BattleResult AiBattleSimulator::RunBattle() {
    std::cout << "  Map: " << config_.map_name() << "\n";
    std::cout << "  Month: " << config_.month() << "\n";
    std::cout << "  Max rounds: " << config_.max_rounds() << "\n";
-    std::cout << "  Random seed: " << config_.random_seed() << "\n";
    std::cout << "  Attacker AI: "
              << net::eagle0::shardok::ai_battle_simulator::AIAlgorithmType_Name(
                         config_.attacker().ai_algorithm())
@@ -293,7 +87,7 @@ BattleResult AiBattleSimulator::RunBattle() {
              << "\n";

    // Create initial game state
-    auto gameState = initialGameState_.has_value() ? *initialGameState_ : CreateInitialGameState();
+    auto gameState = CreateInitialGameState();

    // Get hex map from game state
    const auto* hexMap = gameState->hex_map();
@@ -305,25 +99,17 @@ BattleResult AiBattleSimulator::RunBattle() {
    // Create the engine once for the entire simulation
    ShardokEngine engine(gameSettings_, gameState);

-    int setupCommands = 0;
-    if (!initialGameState_.has_value()) {
-        // Run setup phase
-        std::cout << "\nStarting setup phase...\n";
-        auto setupResult = config_.use_configured_placements()
-                                   ? RunConfiguredSetupPhase(engine)
-                                   : RunSetupPhase(engine, *attackerAI, *defenderAI);
-        if (setupResult.endReason != BattleResult::EndReason::DRAW) {
-            // Game ended during setup (shouldn't happen normally)
-            return setupResult;
-        }
-        setupCommands = setupResult.totalCommands;
-    } else {
-        std::cout << "\nStarting from saved Shardok state; skipping setup phase.\n";
+    // Run setup phase
+    std::cout << "\nStarting setup phase...\n";
+    auto setupResult = RunSetupPhase(engine, *attackerAI, *defenderAI);
+    if (setupResult.endReason != BattleResult::EndReason::DRAW) {
+        // Game ended during setup (shouldn't happen normally)
+        return setupResult;
    }

    // Run battle phase
    std::cout << "\nStarting battle phase...\n";
-    return RunBattlePhase(engine, *attackerAI, *defenderAI, setupCommands);
+    return RunBattlePhase(engine, *attackerAI, *defenderAI);
 }

 GameStateW AiBattleSimulator::CreateInitialGameState() const {
@@ -371,7 +157,7 @@ GameStateW AiBattleSimulator::CreateInitialGameState() const {
        unit.mutate_remaining_action_points(12);
        unit.mutate_hidden(false);
        unit.mutate_fortified(false);
-        unit.mutate_can_flee(GetCanFleeOrDefault(unitConfig));
+        unit.mutate_can_flee(true);
        unit.mutate_can_start_fire(false);
        unit.mutate_can_archery(false);
        unit.mutate_stun_rounds_remaining(0);
@@ -392,17 +178,18 @@ GameStateW AiBattleSimulator::CreateInitialGameState() const {
        // Get battalion type to determine default capacity
        const auto& battalionType = gameSettings_->GetGetter().GetBattalionType(battalionTypeId);
        const double defaultSize = battalionType->capacity;
-        unit.mutate_can_archery(battalionType->alwaysArcheryCapable || HeroCanArchery(unitConfig));
-        unit.mutate_can_start_fire(HeroCanStartFire(unitConfig));

-        battalion.mutate_size(GetBattalionSizeOrDefault(unitConfig.battalion(), defaultSize));
-        battalion.mutate_armament(GetBattalionArmamentOrDefault(unitConfig.battalion()));
-        battalion.mutate_training(GetBattalionTrainingOrDefault(unitConfig.battalion()));
-        battalion.mutate_morale(GetBattalionMoraleOrDefault(unitConfig.battalion()));
+        battalion.mutate_size(GetOrDefault(unitConfig.battalion().size(), defaultSize));
+        battalion.mutate_armament(
+                GetOrDefault(unitConfig.battalion().armament(), DEFAULT_BATTALION_ARMAMENT));
+        battalion.mutate_training(
+                GetOrDefault(unitConfig.battalion().training(), DEFAULT_BATTALION_TRAINING));
+        battalion.mutate_morale(
+                GetOrDefault(unitConfig.battalion().morale(), DEFAULT_BATTALION_MORALE));
        unit.mutable_battalion() = battalion;

        // Hero
-        if (HasAttachedHero(unitConfig.profession())) {
+        if (unitConfig.profession() > 0) {
            unit.mutate_has_attached_hero(true);

            net::eagle0::shardok::storage::fb::Hero hero;
@@ -441,9 +228,6 @@ GameStateW AiBattleSimulator::CreateInitialGameState() const {
        } else {
            unit.mutate_has_attached_hero(false);
        }
-        if (battalion.size() <= 0 && !unit.has_attached_hero()) {
-            unit.mutate_status(net::eagle0::shardok::storage::fb::UnitStatus_RESERVED_SLOT);
-        }

        // Initialize opponent knowledge
        unit.mutable_opponent_knowledge()->Mutate(0, 0);
@@ -466,7 +250,7 @@ GameStateW AiBattleSimulator::CreateInitialGameState() const {
        unit.mutate_remaining_action_points(0);  // Player 1 starts with 0 AP (not their turn yet)
        unit.mutate_hidden(false);
        unit.mutate_fortified(false);
-        unit.mutate_can_flee(GetCanFleeOrDefault(unitConfig));
+        unit.mutate_can_flee(true);
        unit.mutate_can_start_fire(false);
        unit.mutate_can_archery(false);
        unit.mutate_stun_rounds_remaining(0);
@@ -487,17 +271,18 @@ GameStateW AiBattleSimulator::CreateInitialGameState() const {
        // Get battalion type to determine default capacity
        const auto& battalionType = gameSettings_->GetGetter().GetBattalionType(battalionTypeId);
        const double defaultSize = battalionType->capacity;
-        unit.mutate_can_archery(battalionType->alwaysArcheryCapable || HeroCanArchery(unitConfig));
-        unit.mutate_can_start_fire(HeroCanStartFire(unitConfig));

-        battalion.mutate_size(GetBattalionSizeOrDefault(unitConfig.battalion(), defaultSize));
-        battalion.mutate_armament(GetBattalionArmamentOrDefault(unitConfig.battalion()));
-        battalion.mutate_training(GetBattalionTrainingOrDefault(unitConfig.battalion()));
-        battalion.mutate_morale(GetBattalionMoraleOrDefault(unitConfig.battalion()));
+        battalion.mutate_size(GetOrDefault(unitConfig.battalion().size(), defaultSize));
+        battalion.mutate_armament(
+                GetOrDefault(unitConfig.battalion().armament(), DEFAULT_BATTALION_ARMAMENT));
+        battalion.mutate_training(
+                GetOrDefault(unitConfig.battalion().training(), DEFAULT_BATTALION_TRAINING));
+        battalion.mutate_morale(
+                GetOrDefault(unitConfig.battalion().morale(), DEFAULT_BATTALION_MORALE));
        unit.mutable_battalion() = battalion;

        // Hero
-        if (HasAttachedHero(unitConfig.profession())) {
+        if (unitConfig.profession() > 0) {
            unit.mutate_has_attached_hero(true);

            net::eagle0::shardok::storage::fb::Hero hero;
@@ -536,9 +321,6 @@ GameStateW AiBattleSimulator::CreateInitialGameState() const {
        } else {
            unit.mutate_has_attached_hero(false);
        }
-        if (battalion.size() <= 0 && !unit.has_attached_hero()) {
-            unit.mutate_status(net::eagle0::shardok::storage::fb::UnitStatus_RESERVED_SLOT);
-        }

        // Initialize opponent knowledge
        unit.mutable_opponent_knowledge()->Mutate(0, 0);
@@ -566,19 +348,13 @@ std::unique_ptr<ShardokAIClient> AiBattleSimulator::CreateAIClient(
    const bool isDefender = (playerId == DEFENDER_ID);

    AIAlgorithmType algorithmType = ConvertAIAlgorithmType(playerConfig.ai_algorithm());
-    const ScoringCalculatorType scoringType =
-            ConvertScoringCalculatorType(playerConfig.scoring_calculator());
-    const auto mctsConfig = BuildMCTSConfig(playerConfig, config_.random_seed(), playerId);

    return ai_testing_common::AIClientFactory::Create(
            playerId,
            isDefender,
            hexMap,
            gameSettings_->GetGetter(),
-            algorithmType,
-            scoringType,
-            true,
-            mctsConfig);
+            algorithmType);
 }

 BattleResult AiBattleSimulator::RunSetupPhase(
@@ -589,67 +365,7 @@ BattleResult AiBattleSimulator::RunSetupPhase(
        return (playerId == ATTACKER_ID) ? attackerAI : defenderAI;
    };

-    ai_testing_common::PhaseResult phaseResult;
-
-    while (engine.GetCurrentGameState()->status()->state() ==
-           net::eagle0::shardok::storage::fb::GameStatus_::State_SET_UP) {
-        auto currentState = engine.GetCurrentGameState();
-        PlayerId currentPlayer = currentState->current_player();
-
-        auto availableCommands = engine.GetAvailableCommandProtos(currentPlayer, false);
-        if (availableCommands.empty()) { break; }
-
-        ShardokAIClient& activeAI = getAI(currentPlayer);
-        const SideTotals preTotals = CalculateSideTotals(currentState);
-        auto choiceResults = activeAI.ChooseCommandIndex(engine);
-        const auto command = availableCommands.at(choiceResults.chosenIndex);
-
-        const auto randomGenerator = gameSettings_->GetGetter().GetRandomGenerator();
-        engine.PostCommand(currentPlayer, choiceResults.chosenIndex, randomGenerator);
-        phaseResult.commandsExecuted++;
-        const auto postState = engine.GetCurrentGameState();
-        const SideTotals postTotals = CalculateSideTotals(postState);
-
-        if (traceOptions_.output != nullptr) {
-            *traceOptions_.output
-                    << "{"
-                    << "\"run_label\":\"" << JsonEscape(traceOptions_.runLabel) << "\","
-                    << "\"config\":\"" << JsonEscape(traceOptions_.configPath) << "\","
-                    << "\"state_file\":\"" << JsonEscape(traceOptions_.statePath) << "\","
-                    << "\"phase\":\"setup\","
-                    << "\"sequence\":" << phaseResult.commandsExecuted << ","
-                    << "\"round\":" << static_cast<int>(currentState->current_round()) << ","
-                    << "\"player\":" << static_cast<int>(currentPlayer) << ","
-                    << "\"side\":\"" << (currentPlayer == ATTACKER_ID ? "attacker" : "defender")
-                    << "\","
-                    << "\"command_type\":\""
-                    << net::eagle0::shardok::common::CommandType_Name(command.type()) << "\","
-                    << "\"actor_unit\":" << (command.has_actor() ? command.actor().value() : -1)
-                    << ","
-                    << "\"target_row\":" << command.target().row() << ","
-                    << "\"target_column\":" << command.target().column() << ","
-                    << "\"chosen_index\":" << choiceResults.chosenIndex << ","
-                    << "\"available_command_count\":" << choiceResults.availableCommandCount << ","
-                    << "\"depth_achieved\":" << choiceResults.depthAchieved << ","
-                    << "\"commands_evaluated\":" << choiceResults.commandCountEvaluated << ","
-                    << "\"completion_reason\":\""
-                    << CompletionReasonLabel(choiceResults.completionReason) << "\","
-                    << "\"forced_commands\":0,"
-                    << "\"pre_attacker_units\":" << preTotals.attackerUnits << ","
-                    << "\"pre_defender_units\":" << preTotals.defenderUnits << ","
-                    << "\"pre_attacker_troops\":" << preTotals.attackerTroops << ","
-                    << "\"pre_defender_troops\":" << preTotals.defenderTroops << ","
-                    << "\"post_attacker_units\":" << postTotals.attackerUnits << ","
-                    << "\"post_defender_units\":" << postTotals.defenderUnits << ","
-                    << "\"post_attacker_troops\":" << postTotals.attackerTroops << ","
-                    << "\"post_defender_troops\":" << postTotals.defenderTroops << "}\n";
-        }
-
-        if (engine.GameIsOver()) {
-            phaseResult.gameEnded = true;
-            break;
-        }
-    }
+    auto phaseResult = ai_testing_common::GamePhaseRunner::RunSetupPhase(engine, getAI);

    std::cout << "Setup phase complete. Commands executed: " << phaseResult.commandsExecuted
              << "\n";
@@ -672,72 +388,14 @@ BattleResult AiBattleSimulator::RunSetupPhase(
    return result;
 }

-BattleResult AiBattleSimulator::RunConfiguredSetupPhase(ShardokEngine& engine) {
-    auto placePlayer = [&](PlayerId playerId,
-                           const net::eagle0::shardok::ai_battle_simulator::PlayerConfig&
-                                   playerConfig,
-                           int unitIdOffset) {
-        std::vector<UnitPlacementInfo> placements;
-        placements.reserve(playerConfig.units_size());
-        for (int i = 0; i < playerConfig.units_size(); ++i) {
-            const auto& unitConfig = playerConfig.units(i);
-            if (!unitConfig.has_fixed_position()) {
-                throw std::invalid_argument(
-                        "use_configured_placements requires has_fixed_position for every unit");
-            }
-            placements.emplace_back(
-                    static_cast<UnitId>(unitIdOffset + i),
-                    Coords(unitConfig.row(), unitConfig.column()));
-        }
-
-        const auto randomGenerator = gameSettings_->GetGetter().GetRandomGenerator();
-        engine.PostPlacementCommands(playerId, placements, randomGenerator);
-        engine.PostFinishedPlacementCommand(playerId, randomGenerator);
-        if (config_.apply_forced_commands()) {
-            engine.PostWhileCurrentPlayerHasOnlyOneOption(randomGenerator);
-        }
-    };
-
-    placePlayer(ATTACKER_ID, config_.attacker(), 0);
-    if (!engine.GameIsOver()) {
-        placePlayer(DEFENDER_ID, config_.defender(), config_.attacker().units_size());
-    }
-
-    std::cout << "Configured setup complete.\n";
-
-    if (engine.GameIsOver()) {
-        return CreateResultFromGameState(
-                engine.GetCurrentGameState(),
-                0,
-                static_cast<int>(engine.GetUnfilteredHistoryCount()));
-    }
-
-    BattleResult result;
-    result.winner = -1;
-    result.totalRounds = 0;
-    result.totalCommands = static_cast<int>(engine.GetUnfilteredHistoryCount());
-    result.endReason = BattleResult::EndReason::DRAW;
-    result.description = "Configured setup completed";
-    return result;
-}
-
 BattleResult AiBattleSimulator::RunBattlePhase(
        ShardokEngine& engine,
        ShardokAIClient& attackerAI,
-        ShardokAIClient& defenderAI,
-        int setupCommands) {
-    int totalCommands = setupCommands;
-    int attackerCommands = 0;
-    int defenderCommands = 0;
-    int attackerDepthTotal = 0;
-    int defenderDepthTotal = 0;
-    int currentRound = std::max<int>(1, engine.GetCurrentGameState()->current_round());
+        ShardokAIClient& defenderAI) {
+    int totalCommands = 0;
+    int currentRound = 1;

-    const int harnessMaxRounds = config_.max_rounds() > 0
-                                         ? config_.max_rounds()
-                                         : gameSettings_->GetGetter().Backing().max_rounds();
-
-    while (!engine.GameIsOver() && currentRound <= harnessMaxRounds) {
+    while (!engine.GameIsOver() && currentRound <= config_.max_rounds()) {
        auto currentState = engine.GetCurrentGameState();
        PlayerId currentPlayer = currentState->current_player();

@@ -752,66 +410,14 @@ BattleResult AiBattleSimulator::RunBattlePhase(
        ShardokAIClient& activeAI = (currentPlayer == ATTACKER_ID) ? attackerAI : defenderAI;

        // Get AI decision
-        const SideTotals preTotals = CalculateSideTotals(currentState);
        auto choiceResults = activeAI.ChooseCommandIndex(engine);
-        const auto command = availableCommands.at(choiceResults.chosenIndex);

        // Apply command
-        const auto randomGenerator = gameSettings_->GetGetter().GetRandomGenerator();
-        engine.PostCommand(currentPlayer, choiceResults.chosenIndex, randomGenerator);
+        engine.PostCommand(currentPlayer, choiceResults.chosenIndex);
        totalCommands++;
-        int forcedCommands = 0;
-        if (currentPlayer == ATTACKER_ID) {
-            attackerCommands++;
-            attackerDepthTotal += choiceResults.depthAchieved;
-        } else if (currentPlayer == DEFENDER_ID) {
-            defenderCommands++;
-            defenderDepthTotal += choiceResults.depthAchieved;
-        }
-
-        if (config_.apply_forced_commands()) {
-            forcedCommands = engine.PostWhileCurrentPlayerHasOnlyOneOption(randomGenerator);
-            totalCommands += forcedCommands;
-        }

        // Check round progression
        auto newState = engine.GetCurrentGameState();
-        const SideTotals postTotals = CalculateSideTotals(newState);
-        if (traceOptions_.output != nullptr) {
-            const int sequence = totalCommands - forcedCommands;
-            *traceOptions_.output
-                    << "{"
-                    << "\"run_label\":\"" << JsonEscape(traceOptions_.runLabel) << "\","
-                    << "\"config\":\"" << JsonEscape(traceOptions_.configPath) << "\","
-                    << "\"state_file\":\"" << JsonEscape(traceOptions_.statePath) << "\","
-                    << "\"phase\":\"battle\","
-                    << "\"sequence\":" << sequence << ","
-                    << "\"round\":" << static_cast<int>(currentState->current_round()) << ","
-                    << "\"player\":" << static_cast<int>(currentPlayer) << ","
-                    << "\"side\":\"" << (currentPlayer == ATTACKER_ID ? "attacker" : "defender")
-                    << "\","
-                    << "\"command_type\":\""
-                    << net::eagle0::shardok::common::CommandType_Name(command.type()) << "\","
-                    << "\"actor_unit\":" << (command.has_actor() ? command.actor().value() : -1)
-                    << ","
-                    << "\"target_row\":" << command.target().row() << ","
-                    << "\"target_column\":" << command.target().column() << ","
-                    << "\"chosen_index\":" << choiceResults.chosenIndex << ","
-                    << "\"available_command_count\":" << choiceResults.availableCommandCount << ","
-                    << "\"depth_achieved\":" << choiceResults.depthAchieved << ","
-                    << "\"commands_evaluated\":" << choiceResults.commandCountEvaluated << ","
-                    << "\"completion_reason\":\""
-                    << CompletionReasonLabel(choiceResults.completionReason) << "\","
-                    << "\"forced_commands\":" << forcedCommands << ","
-                    << "\"pre_attacker_units\":" << preTotals.attackerUnits << ","
-                    << "\"pre_defender_units\":" << preTotals.defenderUnits << ","
-                    << "\"pre_attacker_troops\":" << preTotals.attackerTroops << ","
-                    << "\"pre_defender_troops\":" << preTotals.defenderTroops << ","
-                    << "\"post_attacker_units\":" << postTotals.attackerUnits << ","
-                    << "\"post_defender_units\":" << postTotals.defenderUnits << ","
-                    << "\"post_attacker_troops\":" << postTotals.attackerTroops << ","
-                    << "\"post_defender_troops\":" << postTotals.defenderTroops << "}\n";
-        }
        if (newState->current_round() > currentRound) {
            std::cout << "Round " << currentRound
                      << " completed. Commands this round: " << totalCommands << "\n";
@@ -822,23 +428,7 @@ BattleResult AiBattleSimulator::RunBattlePhase(
    std::cout << "Battle phase complete. Total rounds: " << currentRound
              << ", Total commands: " << totalCommands << "\n";

-    auto result =
-            CreateResultFromGameState(engine.GetCurrentGameState(), currentRound, totalCommands);
-    if (!engine.GameIsOver() && currentRound > harnessMaxRounds) {
-        result.winner = DEFENDER_ID;
-        result.totalRounds = harnessMaxRounds;
-        result.endReason = BattleResult::EndReason::MAX_ROUNDS_REACHED;
-        result.description = "Defender won by surviving maximum rounds";
-    }
-    result.attackerCommands = attackerCommands;
-    result.defenderCommands = defenderCommands;
-    result.attackerAverageDepth =
-            attackerCommands == 0 ? 0.0
-                                  : static_cast<double>(attackerDepthTotal) / attackerCommands;
-    result.defenderAverageDepth =
-            defenderCommands == 0 ? 0.0
-                                  : static_cast<double>(defenderDepthTotal) / defenderCommands;
-    return result;
+    return CreateResultFromGameState(engine.GetCurrentGameState(), currentRound, totalCommands);
 }

 bool AiBattleSimulator::IsGameOver(const GameStateW& state) const {
@@ -865,10 +455,10 @@ BattleResult AiBattleSimulator::CreateResultFromGameState(
        if (winningIds && winningIds->size() > 0) {
            result.winner = winningIds->Get(0);

+            // Determine how the game ended
            if (result.winner == ATTACKER_ID) {
-                if (status->end_game_condition()->victory_details() ==
-                    net::eagle0::shardok::storage::fb::
-                            VictoryCondition_VICTORY_CONDITION_HOLDS_CRITICAL_TILES) {
+                // Attacker won
+                if (totalRounds >= config_.max_rounds()) {
                    result.endReason = BattleResult::EndReason::CRITICAL_TILES_CONTROLLED;
                    result.description = "Attacker won by controlling critical tiles";
                } else {
@@ -876,9 +466,8 @@ BattleResult AiBattleSimulator::CreateResultFromGameState(
                    result.description = "Attacker won by eliminating all defenders";
                }
            } else if (result.winner == DEFENDER_ID) {
-                if (status->end_game_condition()->victory_details() ==
-                    net::eagle0::shardok::storage::fb::
-                            VictoryCondition_VICTORY_CONDITION_WIN_AFTER_MAX_ROUNDS) {
+                // Defender won
+                if (totalRounds >= config_.max_rounds()) {
                    result.endReason = BattleResult::EndReason::MAX_ROUNDS_REACHED;
                    result.description = "Defender won by surviving maximum rounds";
                } else {
@@ -900,11 +489,10 @@ BattleResult AiBattleSimulator::CreateResultFromGameState(
        result.endReason = BattleResult::EndReason::DRAW;
        result.description = "Game ended in a draw";
    } else {
-        // If the harness stops before the engine posts its max-round status transition, Shardok's
-        // battle rule still awards the battle to the defender.
-        result.winner = DEFENDER_ID;
-        result.endReason = BattleResult::EndReason::MAX_ROUNDS_REACHED;
-        result.description = "Defender won by surviving maximum rounds";
+        // Game not over yet (shouldn't happen)
+        result.winner = -1;
+        result.endReason = BattleResult::EndReason::DRAW;
+        result.description = "Game incomplete";
    }

    // Collect surviving units (for winner, or all units if draw)
@@ -923,14 +511,6 @@ BattleResult AiBattleSimulator::CreateResultFromGameState(

            if (!isAlive) { continue; }

-            if (unit->player_id() == ATTACKER_ID) {
-                result.attackerSurvivingUnits++;
-                result.attackerSurvivingTroops += unit->battalion().size();
-            } else if (unit->player_id() == DEFENDER_ID) {
-                result.defenderSurvivingUnits++;
-                result.defenderSurvivingTroops += unit->battalion().size();
-            }
-
            // For victories, only include winner's units; for draws, include all
            if (result.winner != -1 && unit->player_id() != result.winner) { continue; }

@@ -6,8 +6,6 @@
 #define EAGLE0_AI_BATTLE_SIMULATOR_HPP

 #include <memory>
-#include <optional>
-#include <ostream>
 #include <string>
 #include <vector>

@@ -53,20 +51,6 @@ struct BattleResult {
    // Total number of commands executed
    int totalCommands;

-    // AI-selected commands by player.
-    int attackerCommands = 0;
-    int defenderCommands = 0;
-
-    // Approximate average search depth reported by each AI.
-    double attackerAverageDepth = 0.0;
-    double defenderAverageDepth = 0.0;
-
-    // Surviving units and troop totals by side, useful for scoring A/B tests.
-    int attackerSurvivingUnits = 0;
-    int defenderSurvivingUnits = 0;
-    double attackerSurvivingTroops = 0.0;
-    double defenderSurvivingTroops = 0.0;
-
    // How the game ended
    enum class EndReason {
        LAST_PLAYER_STANDING,       // One player eliminated all opponent units
@@ -83,13 +67,6 @@ struct BattleResult {
    std::vector<SurvivingUnit> survivingUnits;
 };

-struct BattleTraceOptions {
-    std::ostream* output = nullptr;
-    std::string runLabel;
-    std::string configPath;
-    std::string statePath;
-};
-
 class AiBattleSimulator {
 public:
    /**
@@ -100,9 +77,7 @@ public:
     */
    explicit AiBattleSimulator(
            const BattleConfigProto& config,
-            GameSettingsSPtr gameSettings = nullptr,
-            std::optional<GameStateW> initialGameState = std::nullopt,
-            BattleTraceOptions traceOptions = {});
+            GameSettingsSPtr gameSettings = nullptr);

    /**
     * Run the battle from start to completion.
@@ -115,13 +90,9 @@ public:
    [[nodiscard]] BattleResult RunBattle();

 private:
-    friend class AiBattleSimulatorTest_GeneratedUnitsMirrorProductionCapabilityFlags_Test;
-
    // Configuration
    const BattleConfigProto config_;
    GameSettingsSPtr gameSettings_;
-    std::optional<GameStateW> initialGameState_;
-    BattleTraceOptions traceOptions_;

    // Game state builder and helpers
    [[nodiscard]] GameStateW CreateInitialGameState() const;
@@ -133,13 +104,8 @@ private:
    [[nodiscard]] BattleResult
    RunSetupPhase(ShardokEngine& engine, ShardokAIClient& attackerAI, ShardokAIClient& defenderAI);

-    [[nodiscard]] BattleResult RunConfiguredSetupPhase(ShardokEngine& engine);
-
-    [[nodiscard]] BattleResult RunBattlePhase(
-            ShardokEngine& engine,
-            ShardokAIClient& attackerAI,
-            ShardokAIClient& defenderAI,
-            int setupCommands);
+    [[nodiscard]] BattleResult
+    RunBattlePhase(ShardokEngine& engine, ShardokAIClient& attackerAI, ShardokAIClient& defenderAI);

    [[nodiscard]] bool IsGameOver(const GameStateW& state) const;
    [[nodiscard]] BattleResult
@@ -21,7 +21,6 @@ cc_library(
    deps = [
        ":ai_battle_config",
        "//src/main/cpp/net/eagle0/common:filesystem_utils",
-        "//src/main/cpp/net/eagle0/common:random_generator",
        "//src/main/cpp/net/eagle0/common:tsv_parser",
        "//src/main/cpp/net/eagle0/shardok/ai:shardok_ai_client",
        "//src/main/cpp/net/eagle0/shardok/ai_testing_common:ai_client_factory",
@@ -29,7 +28,6 @@ cc_library(
        "//src/main/cpp/net/eagle0/shardok/ai_testing_common:game_settings_factory",
        "//src/main/cpp/net/eagle0/shardok/library:engine",
        "//src/main/cpp/net/eagle0/shardok/library:game_state_w",
-        "//src/main/cpp/net/eagle0/shardok/library:unit_placement_info",
        "//src/main/cpp/net/eagle0/shardok/library/fb_helpers:game_state_helpers",
        "//src/main/cpp/net/eagle0/shardok/library/settings:game_settings",
        "//src/main/cpp/net/eagle0/shardok/util:battalion_type_registrar",
@@ -44,7 +42,7 @@ cc_binary(
    name = "ai_battle_simulator_main",
    srcs = ["ai_battle_simulator_main.cpp"],
    copts = COPTS,
-    data = glob(["benchmarks/*.json"]) + [
+    data = [
        "//src/main/resources/net/eagle0/shardok:battalion_types",
        "//src/main/resources/net/eagle0/shardok:settings",
        "//src/main/resources/net/eagle0/shardok/maps",
@@ -56,26 +54,3 @@ cc_binary(
        "//src/main/cpp/net/eagle0/shardok/library/action_point_distances:fixed_action_point_distances",
    ],
 )
-
-cc_binary(
-    name = "real_battle_config_extractor",
-    srcs = ["real_battle_config_extractor_main.cpp"],
-    copts = COPTS + ["-Wno-deprecated-declarations"],
-    deps = [
-        ":ai_battle_config",
-        "//src/main/protobuf/net/eagle0/eagle/internal:game_state_cc_proto",
-        "//src/main/protobuf/net/eagle0/eagle/internal:shardok_battle_cc_proto",
-    ],
-)
-
-cc_binary(
-    name = "real_battle_smoke_metadata",
-    srcs = ["real_battle_smoke_metadata_main.cpp"],
-    copts = COPTS + ["-Wno-deprecated-declarations"],
-    deps = [
-        "//src/main/flatbuffer/net/eagle0/shardok/storage:game_state_cc_fbs",
-        "//src/main/protobuf/net/eagle0/eagle/internal:shardok_battle_cc_proto",
-        "//src/main/protobuf/net/eagle0/shardok/api:action_result_view_cc_proto",
-        "//src/main/protobuf/net/eagle0/shardok/storage:action_result_cc_proto",
-    ],
-)
@@ -2,78 +2,37 @@
 // AI Battle Simulator - CLI Entry Point
 //

-#include <fcntl.h>
-#include <unistd.h>
-
-#include <cstdio>
 #include <filesystem>
-#include <fstream>
 #include <iostream>
 #include <memory>
-#include <optional>
-#include <sstream>
-#include <stdexcept>
 #include <string>
-#include <vector>

 #include "AiBattleConfig.hpp"
 #include "AiBattleSimulator.hpp"
 #include "src/main/cpp/net/eagle0/common/FilesystemUtils.hpp"

-using shardok::GameStateW;
 using shardok::ai_battle_simulator::AiBattleConfigLoader;
 using shardok::ai_battle_simulator::AiBattleSimulator;
 using shardok::ai_battle_simulator::BattleResult;
-using shardok::ai_battle_simulator::BattleTraceOptions;

 namespace {

-enum class SummaryFormat {
-    HUMAN,
-    CSV,
-    JSON,
-};
-
-class NullBuffer : public std::streambuf {
-protected:
-    int overflow(int c) override { return c; }
-};
-
 void PrintUsage(const char* programName) {
    std::cout << "AI Battle Simulator - AI vs AI Battle Testing Tool\n"
              << "\n"
              << "Usage:\n"
              << "  " << programName << " --config=<path>    Run battle from JSON config file\n"
-              << "  " << programName << " --config=<path> --config=<path> --summary=csv --quiet\n"
-              << "  " << programName
-              << " --config=<path> --state-file=<path> --summary=csv --quiet\n"
-              << "  " << programName
-              << " --config=<path> --trace-jsonl=<path> --run-label=baseline\n"
              << "  " << programName << " --generate-config  Generate sample config to stdout\n"
              << "  " << programName
              << " --generate-config --output=<path>  Generate sample config to file\n"
              << "  " << programName << " --help             Show this help message\n"
              << "\n"
-              << "Options:\n"
-              << "  --summary=human|csv|json  Output format for completed battle summaries\n"
-              << "  --quiet                   Suppress simulator progress logs\n"
-              << "  --trace-jsonl=<path>      Write one JSON object per AI-selected command\n"
-              << "  --run-label=<label>       Label included in trace rows\n"
-              << "\n"
              << "Examples:\n"
              << "  # Generate sample config\n"
              << "  " << programName << " --generate-config > my_battle.json\n"
              << "\n"
              << "  # Run battle from config\n"
              << "  " << programName << " --config=my_battle.json\n"
-              << "\n"
-              << "  # Run a compact benchmark suite\n"
-              << "  " << programName
-              << " --config=baseline.json --config=candidate.json --summary=csv --quiet\n"
-              << "\n"
-              << "  # Run an AI config from a saved Shardok GameState flatbuffer\n"
-              << "  " << programName
-              << " --config=ai_config.json --state-file=state.fb --summary=csv --quiet\n"
              << "\n";
 }

@@ -96,50 +55,23 @@ void GenerateConfigFile(const std::string& outputPath) {
    }
 }

-std::string CsvEscape(const std::string& value) {
-    bool mustQuote = false;
-    std::string escaped;
-    escaped.reserve(value.size());
-    for (const char c : value) {
-        if (c == '"' || c == ',' || c == '\n' || c == '\r') { mustQuote = true; }
-        if (c == '"') { escaped.push_back('"'); }
-        escaped.push_back(c);
+void RunBattleFromConfig(const std::string& configPath) {
+    // Load config
+    std::cout << "Loading config from: " << configPath << "\n";
+    auto config = AiBattleConfigLoader::LoadFromJsonFile(configPath);
+
+    if (!config) {
+        std::cerr << "Error: Failed to load config file\n";
+        std::exit(1);
    }
-    if (!mustQuote) { return escaped; }
-    return "\"" + escaped + "\"";
-}

-std::string JsonEscape(const std::string& value) {
-    std::ostringstream out;
-    for (const char c : value) {
-        switch (c) {
-            case '"': out << "\\\""; break;
-            case '\\': out << "\\\\"; break;
-            case '\n': out << "\\n"; break;
-            case '\r': out << "\\r"; break;
-            case '\t': out << "\\t"; break;
-            default: out << c; break;
-        }
-    }
-    return out.str();
-}
+    // Create simulator
+    AiBattleSimulator simulator(*config);

-std::string WinnerLabel(int winner) {
-    if (winner == 0) { return "attacker"; }
-    if (winner == 1) { return "defender"; }
-    return "draw";
-}
+    // Run battle
+    auto result = simulator.RunBattle();

-std::string EndReasonLabel(BattleResult::EndReason endReason) {
-    switch (endReason) {
-        case BattleResult::EndReason::LAST_PLAYER_STANDING: return "last_player_standing";
-        case BattleResult::EndReason::CRITICAL_TILES_CONTROLLED: return "critical_tiles_controlled";
-        case BattleResult::EndReason::MAX_ROUNDS_REACHED: return "max_rounds_reached";
-        case BattleResult::EndReason::DRAW: return "draw";
-    }
-}
-
-void PrintHumanSummary(const BattleResult& result) {
+    // Print results
    std::cout << "\n";
    std::cout << "=================================\n";
    std::cout << "Battle Complete\n";
@@ -155,17 +87,10 @@ void PrintHumanSummary(const BattleResult& result) {
    std::cout << "Result: " << result.description << "\n";
    std::cout << "Total Rounds: " << result.totalRounds << "\n";
    std::cout << "Total Commands: " << result.totalCommands << "\n";
-    std::cout << "Attacker AI Commands: " << result.attackerCommands << "\n";
-    std::cout << "Defender AI Commands: " << result.defenderCommands << "\n";
-    std::cout << "Attacker Avg Depth: " << result.attackerAverageDepth << "\n";
-    std::cout << "Defender Avg Depth: " << result.defenderAverageDepth << "\n";
-    std::cout << "Attacker Survivors: " << result.attackerSurvivingUnits << " units, "
-              << result.attackerSurvivingTroops << " troops\n";
-    std::cout << "Defender Survivors: " << result.defenderSurvivingUnits << " units, "
-              << result.defenderSurvivingTroops << " troops\n";

+    // Print surviving units
    if (!result.survivingUnits.empty()) {
-        std::cout << "\nWinning/Draw Surviving Units (" << result.survivingUnits.size() << "):\n";
+        std::cout << "\nSurviving Units (" << result.survivingUnits.size() << "):\n";
        for (const auto& unit : result.survivingUnits) {
            std::cout << "  Unit " << unit.unitId << ": " << unit.battalionTypeName << " ("
                      << static_cast<int>(unit.battalionSize) << " troops)";
@@ -177,143 +102,6 @@ void PrintHumanSummary(const BattleResult& result) {
    std::cout << "=================================\n";
 }

-void PrintCsvHeader() {
-    std::cout << "config,state_file,winner,end_reason,result,total_rounds,total_commands,"
-              << "attacker_ai_commands,defender_ai_commands,attacker_avg_depth,"
-              << "defender_avg_depth,attacker_surviving_units,defender_surviving_units,"
-              << "attacker_surviving_troops,defender_surviving_troops\n";
-}
-
-void PrintCsvRow(
-        const std::string& configPath,
-        const std::string& statePath,
-        const BattleResult& result) {
-    std::cout << CsvEscape(configPath) << "," << CsvEscape(statePath) << ","
-              << WinnerLabel(result.winner) << "," << EndReasonLabel(result.endReason) << ","
-              << CsvEscape(result.description) << "," << result.totalRounds << ","
-              << result.totalCommands << "," << result.attackerCommands << ","
-              << result.defenderCommands << "," << result.attackerAverageDepth << ","
-              << result.defenderAverageDepth << "," << result.attackerSurvivingUnits << ","
-              << result.defenderSurvivingUnits << "," << result.attackerSurvivingTroops << ","
-              << result.defenderSurvivingTroops << "\n";
-}
-
-void PrintJsonSummary(
-        const std::string& configPath,
-        const std::string& statePath,
-        const BattleResult& result) {
-    std::cout << "{"
-              << "\"config\":\"" << JsonEscape(configPath) << "\","
-              << "\"state_file\":\"" << JsonEscape(statePath) << "\","
-              << "\"winner\":\"" << WinnerLabel(result.winner) << "\","
-              << "\"end_reason\":\"" << EndReasonLabel(result.endReason) << "\","
-              << "\"result\":\"" << JsonEscape(result.description) << "\","
-              << "\"total_rounds\":" << result.totalRounds << ","
-              << "\"total_commands\":" << result.totalCommands << ","
-              << "\"attacker_ai_commands\":" << result.attackerCommands << ","
-              << "\"defender_ai_commands\":" << result.defenderCommands << ","
-              << "\"attacker_avg_depth\":" << result.attackerAverageDepth << ","
-              << "\"defender_avg_depth\":" << result.defenderAverageDepth << ","
-              << "\"attacker_surviving_units\":" << result.attackerSurvivingUnits << ","
-              << "\"defender_surviving_units\":" << result.defenderSurvivingUnits << ","
-              << "\"attacker_surviving_troops\":" << result.attackerSurvivingTroops << ","
-              << "\"defender_surviving_troops\":" << result.defenderSurvivingTroops << "}\n";
-}
-
-std::unique_ptr<shardok::ai_battle_simulator::BattleConfigProto> LoadConfig(
-        const std::string& configPath) {
-    if (configPath.empty()) { return AiBattleConfigLoader::CreateDefaultPerfConfig(); }
-    return AiBattleConfigLoader::LoadFromJsonFile(configPath);
-}
-
-GameStateW LoadGameStateFromFile(const std::string& statePath) {
-    if (statePath.empty()) { return GameStateW{}; }
-    auto bytes = FilesystemUtils::LoadFromPath(statePath);
-    if (bytes.empty()) { throw std::runtime_error("Failed to load state file: " + statePath); }
-    return GameStateW::FromByteVector(bytes);
-}
-
-BattleResult RunBattle(
-        const std::string& configPath,
-        const std::string& statePath,
-        bool quiet,
-        const std::string& tracePath,
-        const std::string& runLabel) {
-    std::ofstream traceFile;
-    BattleTraceOptions traceOptions;
-    if (!tracePath.empty()) {
-        traceFile.open(tracePath);
-        if (!traceFile.is_open()) {
-            throw std::runtime_error("Failed to open trace file: " + tracePath);
-        }
-        traceOptions.output = &traceFile;
-        traceOptions.runLabel = runLabel;
-        traceOptions.configPath = configPath;
-        traceOptions.statePath = statePath;
-    }
-
-    NullBuffer nullBuffer;
-    std::ostream nullStream(&nullBuffer);
-    auto* originalCout = std::cout.rdbuf();
-    auto* originalCerr = std::cerr.rdbuf();
-    int stdoutCopy = -1;
-    int stderrCopy = -1;
-    if (quiet) {
-        std::cout.rdbuf(nullStream.rdbuf());
-        std::cerr.rdbuf(nullStream.rdbuf());
-        std::fflush(stdout);
-        std::fflush(stderr);
-        stdoutCopy = dup(STDOUT_FILENO);
-        stderrCopy = dup(STDERR_FILENO);
-        const int nullFd = open("/dev/null", O_WRONLY);
-        if (nullFd >= 0) {
-            dup2(nullFd, STDOUT_FILENO);
-            dup2(nullFd, STDERR_FILENO);
-            close(nullFd);
-        }
-    }
-
-    auto restoreOutput = [&]() {
-        if (!quiet) { return; }
-        std::fflush(stdout);
-        std::fflush(stderr);
-        if (stdoutCopy >= 0) {
-            dup2(stdoutCopy, STDOUT_FILENO);
-            close(stdoutCopy);
-        }
-        if (stderrCopy >= 0) {
-            dup2(stderrCopy, STDERR_FILENO);
-            close(stderrCopy);
-        }
-        std::cout.rdbuf(originalCout);
-        std::cerr.rdbuf(originalCerr);
-    };
-
-    // Load config
-    if (!configPath.empty()) { std::cout << "Loading config from: " << configPath << "\n"; }
-    auto config = LoadConfig(configPath);
-
-    if (!config) {
-        restoreOutput();
-        std::cerr << "Error: Failed to load config file\n";
-        std::exit(1);
-    }
-
-    // Create simulator
-    std::optional<GameStateW> initialGameState;
-    if (!statePath.empty()) {
-        std::cout << "Loading saved state from: " << statePath << "\n";
-        initialGameState = LoadGameStateFromFile(statePath);
-    }
-    AiBattleSimulator simulator(*config, nullptr, std::move(initialGameState), traceOptions);
-
-    // Run battle
-    auto result = simulator.RunBattle();
-
-    restoreOutput();
-    return result;
-}
-
 }  // namespace

 int main(int argc, char* argv[]) {
@@ -326,14 +114,9 @@ int main(int argc, char* argv[]) {
            return 1;
        }

-        std::vector<std::string> configPaths;
-        std::vector<std::string> statePaths;
+        std::string configPath;
        std::string outputPath;
-        std::string tracePath;
-        std::string runLabel;
        bool generateConfig = false;
-        bool quiet = false;
-        SummaryFormat summaryFormat = SummaryFormat::HUMAN;

        // Parse command line arguments
        for (int i = 1; i < argc; ++i) {
@@ -345,30 +128,9 @@ int main(int argc, char* argv[]) {
            } else if (arg == "--generate-config") {
                generateConfig = true;
            } else if (arg.starts_with("--config=")) {
-                configPaths.push_back(arg.substr(9));
-            } else if (arg.starts_with("--state-file=")) {
-                statePaths.push_back(arg.substr(13));
+                configPath = arg.substr(9);
            } else if (arg.starts_with("--output=")) {
                outputPath = arg.substr(9);
-            } else if (arg.starts_with("--trace-jsonl=")) {
-                tracePath = arg.substr(14);
-            } else if (arg.starts_with("--run-label=")) {
-                runLabel = arg.substr(12);
-            } else if (arg == "--quiet") {
-                quiet = true;
-            } else if (arg.starts_with("--summary=")) {
-                const auto format = arg.substr(10);
-                if (format == "human") {
-                    summaryFormat = SummaryFormat::HUMAN;
-                } else if (format == "csv") {
-                    summaryFormat = SummaryFormat::CSV;
-                } else if (format == "json") {
-                    summaryFormat = SummaryFormat::JSON;
-                } else {
-                    std::cerr << "Error: Unknown summary format: " << format << "\n";
-                    PrintUsage(argv[0]);
-                    return 1;
-                }
            } else {
                std::cerr << "Error: Unknown argument: " << arg << "\n";
                PrintUsage(argv[0]);
@@ -379,25 +141,10 @@ int main(int argc, char* argv[]) {
        // Execute appropriate action
        if (generateConfig) {
            GenerateConfigFile(outputPath);
-        } else if (!configPaths.empty() || !statePaths.empty()) {
-            if (summaryFormat == SummaryFormat::CSV) { PrintCsvHeader(); }
-            if (configPaths.empty()) { configPaths.emplace_back(""); }
-            if (statePaths.empty()) { statePaths.emplace_back(""); }
-            for (const auto& configPath : configPaths) {
-                for (const auto& statePath : statePaths) {
-                    const auto result =
-                            RunBattle(configPath, statePath, quiet, tracePath, runLabel);
-                    switch (summaryFormat) {
-                        case SummaryFormat::HUMAN: PrintHumanSummary(result); break;
-                        case SummaryFormat::CSV: PrintCsvRow(configPath, statePath, result); break;
-                        case SummaryFormat::JSON:
-                            PrintJsonSummary(configPath, statePath, result);
-                            break;
-                    }
-                }
-            }
+        } else if (!configPath.empty()) {
+            RunBattleFromConfig(configPath);
        } else {
-            std::cerr << "Error: --generate-config, --config, or --state-file must be specified\n";
+            std::cerr << "Error: Either --generate-config or --config must be specified\n";
            PrintUsage(argv[0]);
            return 1;
        }
@@ -1,307 +0,0 @@
-#!/usr/bin/env python3
-"""Compare paired Shardok AI simulator traces for one or more battles."""
-
-from __future__ import annotations
-
-import argparse
-import json
-from collections import Counter, defaultdict
-from dataclasses import dataclass
-from pathlib import Path
-from typing import Iterable
-
-
-STALL_COMMANDS = {
-    "END_TURN_COMMAND",
-    "FLEE_COMMAND",
-    "MOVE_COMMAND",
-    "UNIT_REST_COMMAND",
-}
-
-
-@dataclass(frozen=True)
-class TraceEvent:
-    run_label: str
-    phase: str
-    sequence: int
-    round: int
-    side: str
-    player: int
-    command_type: str
-    actor_unit: int
-    available_command_count: int
-    depth_achieved: int
-    completion_reason: str
-    forced_commands: int
-    pre_attacker_troops: int
-    pre_defender_troops: int
-    post_attacker_troops: int
-    post_defender_troops: int
-
-    @classmethod
-    def from_json(cls, row: dict[str, object]) -> TraceEvent:
-        return cls(
-            run_label=str(row["run_label"]),
-            phase=str(row["phase"]),
-            sequence=int(row["sequence"]),
-            round=int(row["round"]),
-            side=str(row["side"]),
-            player=int(row["player"]),
-            command_type=str(row["command_type"]),
-            actor_unit=int(row["actor_unit"]),
-            available_command_count=int(row["available_command_count"]),
-            depth_achieved=int(row["depth_achieved"]),
-            completion_reason=str(row["completion_reason"]),
-            forced_commands=int(row["forced_commands"]),
-            pre_attacker_troops=int(row["pre_attacker_troops"]),
-            pre_defender_troops=int(row["pre_defender_troops"]),
-            post_attacker_troops=int(row["post_attacker_troops"]),
-            post_defender_troops=int(row["post_defender_troops"]),
-        )
-
-    @property
-    def attacker_delta(self) -> int:
-        return self.post_attacker_troops - self.pre_attacker_troops
-
-    @property
-    def defender_delta(self) -> int:
-        return self.post_defender_troops - self.pre_defender_troops
-
-    @property
-    def did_damage(self) -> bool:
-        return self.attacker_delta < 0 or self.defender_delta < 0
-
-
-@dataclass(frozen=True)
-class RoundSummary:
-    battle_round: int
-    attacker_commands: int
-    defender_commands: int
-    attacker_damage: int
-    defender_damage: int
-    attacker_command_counts: Counter[str]
-    defender_command_counts: Counter[str]
-
-
-@dataclass(frozen=True)
-class RunSummary:
-    label: str
-    final_attacker_troops: int
-    final_defender_troops: int
-    command_counts: Counter[str]
-    side_command_counts: dict[str, Counter[str]]
-    side_damage_done: dict[str, int]
-    no_damage_streaks: list[list[TraceEvent]]
-    rounds: dict[int, RoundSummary]
-
-
-def load_trace(path: Path) -> list[TraceEvent]:
-    events = []
-    with path.open() as f:
-        for line in f:
-            line = line.strip()
-            if line:
-                events.append(TraceEvent.from_json(json.loads(line)))
-    return events
-
-
-def battle_events(events: Iterable[TraceEvent]) -> list[TraceEvent]:
-    return [event for event in events if event.phase == "battle"]
-
-
-def side_damage(event: TraceEvent) -> int:
-    if event.side == "attacker":
-        return max(0, -event.defender_delta)
-    return max(0, -event.attacker_delta)
-
-
-def longest_no_damage_streaks(events: list[TraceEvent], limit: int) -> list[list[TraceEvent]]:
-    streaks = []
-    current = []
-    for event in events:
-        if event.command_type in STALL_COMMANDS and not event.did_damage:
-            current.append(event)
-            continue
-        if current:
-            streaks.append(current)
-            current = []
-    if current:
-        streaks.append(current)
-    streaks.sort(key=len, reverse=True)
-    return streaks[:limit]
-
-
-def summarize_run(label: str, events: list[TraceEvent], streak_limit: int) -> RunSummary:
-    battle = battle_events(events)
-    if not battle:
-        raise RuntimeError(f"No battle events in {label}")
-
-    command_counts = Counter(event.command_type for event in battle)
-    side_command_counts: dict[str, Counter[str]] = {
-        "attacker": Counter(),
-        "defender": Counter(),
-    }
-    side_damage_done = {
-        "attacker": 0,
-        "defender": 0,
-    }
-    per_round: dict[int, list[TraceEvent]] = defaultdict(list)
-    for event in battle:
-        side_command_counts[event.side][event.command_type] += 1
-        side_damage_done[event.side] += side_damage(event)
-        per_round[event.round].append(event)
-
-    rounds = {}
-    for battle_round, round_events in sorted(per_round.items()):
-        attacker_events = [event for event in round_events if event.side == "attacker"]
-        defender_events = [event for event in round_events if event.side == "defender"]
-        rounds[battle_round] = RoundSummary(
-            battle_round=battle_round,
-            attacker_commands=len(attacker_events),
-            defender_commands=len(defender_events),
-            attacker_damage=sum(side_damage(event) for event in attacker_events),
-            defender_damage=sum(side_damage(event) for event in defender_events),
-            attacker_command_counts=Counter(event.command_type for event in attacker_events),
-            defender_command_counts=Counter(event.command_type for event in defender_events),
-        )
-
-    final = battle[-1]
-    return RunSummary(
-        label=label,
-        final_attacker_troops=final.post_attacker_troops,
-        final_defender_troops=final.post_defender_troops,
-        command_counts=command_counts,
-        side_command_counts=side_command_counts,
-        side_damage_done=side_damage_done,
-        no_damage_streaks=longest_no_damage_streaks(battle, streak_limit),
-        rounds=rounds,
-    )
-
-
-def command_counts_text(counts: Counter[str]) -> str:
-    if not counts:
-        return ""
-    return ", ".join(f"{name.removesuffix('_COMMAND')} {count}" for name, count in counts.most_common())
-
-
-def format_streak(streak: list[TraceEvent]) -> str:
-    first = streak[0]
-    last = streak[-1]
-    command_counts = Counter(event.command_type for event in streak)
-    return (
-        f"rounds {first.round}-{last.round}, seq {first.sequence}-{last.sequence}, "
-        f"{len(streak)} commands: {command_counts_text(command_counts)}"
-    )
-
-
-def write_battle_report(
-    lines: list[str],
-    battle: int,
-    baseline: RunSummary,
-    candidate: RunSummary,
-) -> None:
-    lines.extend(
-        [
-            f"## Battle {battle}",
-            "",
-            "| Run | Final troops | Attacker damage | Defender damage | Attacker commands | Defender commands |",
-            "| --- | ---: | ---: | ---: | --- | --- |",
-        ]
-    )
-    for summary in [baseline, candidate]:
-        lines.append(
-            f"| {summary.label} | {summary.final_attacker_troops} vs {summary.final_defender_troops} | "
-            f"{summary.side_damage_done['attacker']} | {summary.side_damage_done['defender']} | "
-            f"{command_counts_text(summary.side_command_counts['attacker'])} | "
-            f"{command_counts_text(summary.side_command_counts['defender'])} |"
-        )
-
-    lines.extend(
-        [
-            "",
-            "### Per-Round Differences",
-            "",
-            "| Round | Baseline attacker | Candidate attacker | Baseline defender | Candidate defender |",
-            "| ---: | --- | --- | --- | --- |",
-        ]
-    )
-    all_rounds = sorted(set(baseline.rounds) | set(candidate.rounds))
-    for battle_round in all_rounds:
-        baseline_round = baseline.rounds.get(battle_round)
-        candidate_round = candidate.rounds.get(battle_round)
-        baseline_attacker = round_side_text(baseline_round, "attacker")
-        candidate_attacker = round_side_text(candidate_round, "attacker")
-        baseline_defender = round_side_text(baseline_round, "defender")
-        candidate_defender = round_side_text(candidate_round, "defender")
-        if baseline_attacker == candidate_attacker and baseline_defender == candidate_defender:
-            continue
-        lines.append(
-            f"| {battle_round} | {baseline_attacker} | {candidate_attacker} | "
-            f"{baseline_defender} | {candidate_defender} |"
-        )
-
-    lines.extend(
-        [
-            "",
-            "### Longest No-Damage Streaks",
-            "",
-            "| Run | Streak |",
-            "| --- | --- |",
-        ]
-    )
-    for summary in [baseline, candidate]:
-        if not summary.no_damage_streaks:
-            lines.append(f"| {summary.label} | _none_ |")
-            continue
-        for streak in summary.no_damage_streaks:
-            lines.append(f"| {summary.label} | {format_streak(streak)} |")
-    lines.append("")
-
-
-def round_side_text(summary: RoundSummary | None, side: str) -> str:
-    if summary is None:
-        return "_none_"
-    if side == "attacker":
-        damage = summary.attacker_damage
-        counts = summary.attacker_command_counts
-        commands = summary.attacker_commands
-    else:
-        damage = summary.defender_damage
-        counts = summary.defender_command_counts
-        commands = summary.defender_commands
-    return f"{commands} cmds, {damage} dmg ({command_counts_text(counts)})"
-
-
-def parse_battles(value: str) -> list[int]:
-    return [int(item.strip()) for item in value.split(",") if item.strip()]
-
-
-def main() -> None:
-    parser = argparse.ArgumentParser()
-    parser.add_argument("--experiment-dir", type=Path, required=True)
-    parser.add_argument("--battles", required=True, help="Comma-separated action_seq values")
-    parser.add_argument("--out", type=Path, required=True)
-    parser.add_argument("--streak-limit", type=int, default=4)
-    args = parser.parse_args()
-
-    lines = [
-        "# Trace Comparison",
-        "",
-        f"- Experiment directory: `{args.experiment_dir}`",
-        f"- Battles: `{args.battles}`",
-        "",
-    ]
-    for battle in parse_battles(args.battles):
-        trace_dir = args.experiment_dir / "traces"
-        baseline_events = load_trace(trace_dir / f"{battle}_baseline.jsonl")
-        candidate_events = load_trace(trace_dir / f"{battle}_candidate.jsonl")
-        baseline = summarize_run("baseline", baseline_events, args.streak_limit)
-        candidate = summarize_run("candidate", candidate_events, args.streak_limit)
-        write_battle_report(lines, battle, baseline, candidate)
-
-    args.out.parent.mkdir(parents=True, exist_ok=True)
-    args.out.write_text("\n".join(lines))
-
-
-if __name__ == "__main__":
-    main()
@@ -1,32 +0,0 @@
-action_seq,shardok_game_id,attacker_factions,defender_factions,real_winner,sim_winner,winner_match,real_round,sim_round,real_attacker_troops,real_defender_troops,sim_attacker_troops,sim_defender_troops
-306,389ffb1901903118_1_b3e7d8ab,1,2,attacker,attacker,true,0,19,1454,0,1388,0
-307,389ffb1901903118_2_bf499d3c,2,1,defender,defender,true,0,16,0,699,0,648
-399,389ffb1901903118_3_cd9141a,2,1,defender,defender,true,0,6,0,1310,0,1315
-631,389ffb1901903118_4_6e7b9b,1,2,defender,defender,true,0,5,0,3766,0,3654
-1326,389ffb1901903118_5_8a57255a,9,6,attacker,attacker,true,0,11,1715,0,1701,0
-2437,389ffb1901903118_7_5cf74e0,2,1,attacker,attacker,true,0,16,3146,0,3247,0
-2584,389ffb1901903118_8_41ce0ebd,1,2,defender,defender,true,0,31,0,2028,0,168
-2654,389ffb1901903118_9_eb43d421,2,1,defender,defender,true,0,28,0,335,0,467
-3536,389ffb1901903118_a_950de1c1,1,2,attacker,attacker,true,0,12,2172,0,2528,0
-3701,389ffb1901903118_b_5d1d76e0,8,2,defender,defender,true,0,30,0,2819,0,1348
-4298,389ffb1901903118_c_7bdb54dd,2,1,defender,defender,true,0,3,0,3514,0,1977
-4576,389ffb1901903118_d_689ee2fa,2,1,defender,defender,true,0,4,0,2028,0,1986
-6014,389ffb1901903118_e_89c90dc6,2,8,defender,defender,true,0,6,0,1942,0,2252
-6248,389ffb1901903118_f_3a9d9136,7,9,attacker,attacker,true,0,25,520,0,676,0
-6705,389ffb1901903118_11_e0611233,1,2,defender,defender,true,0,31,0,3316,0,535
-7212,389ffb1901903118_12_84a36154,2,8,defender,defender,true,0,8,0,1253,0,1273
-8217,389ffb1901903118_13_2e98b1cf,8,2,attacker,attacker,true,0,22,920,0,2167,0
-8394,389ffb1901903118_14_b5570863,2,8,attacker,attacker,true,0,9,2449,0,3186,0
-9975,389ffb1901903118_16_3d6ec341,2,1,defender,defender,true,0,4,0,7420,0,7840
-10229,389ffb1901903118_17_3ba689e3,5,7,attacker,attacker,true,0,23,2135,0,1816,0
-10631,389ffb1901903118_18_9e463db3,1,2,attacker,attacker,true,0,1,4000,0,4000,0
-11011,389ffb1901903118_19_239725d3,2,1,defender,defender,true,0,5,0,3532,0,3658
-11012,389ffb1901903118_1a_16e1ebf,9,6,defender,defender,true,0,30,0,846,0,1347
-11297,389ffb1901903118_1c_529ba6e5,1,2,attacker,attacker,true,0,1,2400,0,2400,0
-11469,389ffb1901903118_1d_49993012,5,9,defender,defender,true,0,8,0,787,0,1489
-11470,389ffb1901903118_1e_fd8c60a8,7,5,attacker,attacker,true,0,17,3220,0,3425,0
-11565,389ffb1901903118_1f_53309484,5,7,defender,defender,true,0,31,190,3420,78,3600
-12136,389ffb1901903118_24_f0d61a5a,9,7,attacker,attacker,true,0,21,1829,0,1581,0
-12137,389ffb1901903118_25_a19c4d87,2,1,defender,defender,true,0,8,0,1894,0,1892
-12765,389ffb1901903118_27_c008989b,6,9,defender,defender,true,0,11,0,1773,0,1763
-13752,389ffb1901903118_28_b33e2868,2,1,defender,defender,true,0,7,0,943,0,932
@@ -1,45 +0,0 @@
-# AI-vs-AI Real Battle Smoke Test
-
-This smoke test reruns saved battles where neither side used faction `3` or `4`,
-then compares STANDARD-vs-STANDARD winner side with the winner side recorded by
-the original Shardok battle. It is intentionally a broad similarity check rather
-than an exact replay assertion.
-
- Checked battles: 31
- Winner-side matches: 31
- Match rate: 100%
- Required match rate: 75%
-
-| Action seq | Factions | Real winner | Sim winner | Real troops | Sim troops |
-| ---: | --- | --- | --- | ---: | ---: |
-| 306 | 1 vs 2 | attacker | attacker | 1454 vs 0 | 1388 vs 0 |
-| 307 | 2 vs 1 | defender | defender | 0 vs 699 | 0 vs 648 |
-| 399 | 2 vs 1 | defender | defender | 0 vs 1310 | 0 vs 1315 |
-| 631 | 1 vs 2 | defender | defender | 0 vs 3766 | 0 vs 3654 |
-| 1326 | 9 vs 6 | attacker | attacker | 1715 vs 0 | 1701 vs 0 |
-| 2437 | 2 vs 1 | attacker | attacker | 3146 vs 0 | 3247 vs 0 |
-| 2584 | 1 vs 2 | defender | defender | 0 vs 2028 | 0 vs 168 |
-| 2654 | 2 vs 1 | defender | defender | 0 vs 335 | 0 vs 467 |
-| 3536 | 1 vs 2 | attacker | attacker | 2172 vs 0 | 2528 vs 0 |
-| 3701 | 8 vs 2 | defender | defender | 0 vs 2819 | 0 vs 1348 |
-| 4298 | 2 vs 1 | defender | defender | 0 vs 3514 | 0 vs 1977 |
-| 4576 | 2 vs 1 | defender | defender | 0 vs 2028 | 0 vs 1986 |
-| 6014 | 2 vs 8 | defender | defender | 0 vs 1942 | 0 vs 2252 |
-| 6248 | 7 vs 9 | attacker | attacker | 520 vs 0 | 676 vs 0 |
-| 6705 | 1 vs 2 | defender | defender | 0 vs 3316 | 0 vs 535 |
-| 7212 | 2 vs 8 | defender | defender | 0 vs 1253 | 0 vs 1273 |
-| 8217 | 8 vs 2 | attacker | attacker | 920 vs 0 | 2167 vs 0 |
-| 8394 | 2 vs 8 | attacker | attacker | 2449 vs 0 | 3186 vs 0 |
-| 9975 | 2 vs 1 | defender | defender | 0 vs 7420 | 0 vs 7840 |
-| 10229 | 5 vs 7 | attacker | attacker | 2135 vs 0 | 1816 vs 0 |
-| 10631 | 1 vs 2 | attacker | attacker | 4000 vs 0 | 4000 vs 0 |
-| 11011 | 2 vs 1 | defender | defender | 0 vs 3532 | 0 vs 3658 |
-| 11012 | 9 vs 6 | defender | defender | 0 vs 846 | 0 vs 1347 |
-| 11297 | 1 vs 2 | attacker | attacker | 2400 vs 0 | 2400 vs 0 |
-| 11469 | 5 vs 9 | defender | defender | 0 vs 787 | 0 vs 1489 |
-| 11470 | 7 vs 5 | attacker | attacker | 3220 vs 0 | 3425 vs 0 |
-| 11565 | 5 vs 7 | defender | defender | 190 vs 3420 | 78 vs 3600 |
-| 12136 | 9 vs 7 | attacker | attacker | 1829 vs 0 | 1581 vs 0 |
-| 12137 | 2 vs 1 | defender | defender | 0 vs 1894 | 0 vs 1892 |
-| 12765 | 6 vs 9 | defender | defender | 0 vs 1773 | 0 vs 1763 |
-| 13752 | 2 vs 1 | defender | defender | 0 vs 943 | 0 vs 932 |
@@ -1,41 +0,0 @@
-variant,action_seq,config,winner,end_reason,result,total_rounds,total_commands,attacker_ai_commands,defender_ai_commands,attacker_avg_depth,defender_avg_depth,attacker_surviving_units,defender_surviving_units,attacker_surviving_troops,defender_surviving_troops,elapsed_seconds
-expected_impact_archery_standard_baseline,306,/private/tmp/eagle0-standard-baselines/configs/306_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,19,99,41,32,4.04878,3.65625,2,0,1388,0,
-expected_impact_archery_standard_baseline,307,/private/tmp/eagle0-standard-baselines/configs/307_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,16,75,37,15,3.54054,2.86667,0,1,0,648,
-expected_impact_archery_standard_baseline,399,/private/tmp/eagle0-standard-baselines/configs/399_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,6,30,7,14,3,3.85714,0,3,0,1315,
-expected_impact_archery_standard_baseline,631,/private/tmp/eagle0-standard-baselines/configs/631_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,5,38,12,14,3.58333,4.14286,0,5,0,3654,
-expected_impact_archery_standard_baseline,1326,/private/tmp/eagle0-standard-baselines/configs/1326_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,11,72,34,19,3.26471,3.63158,2,0,1701,0,
-expected_impact_archery_standard_baseline,1472,/private/tmp/eagle0-standard-baselines/configs/1472_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,31,143,59,67,3.25424,3.46269,0,2,0,944,
-expected_impact_archery_standard_baseline,2437,/private/tmp/eagle0-standard-baselines/configs/2437_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,16,148,89,43,3.37079,3.32558,5,0,3247,0,
-expected_impact_archery_standard_baseline,2584,/private/tmp/eagle0-standard-baselines/configs/2584_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,31,190,48,85,3.45833,3.16471,0,1,0,168,
-expected_impact_archery_standard_baseline,2654,/private/tmp/eagle0-standard-baselines/configs/2654_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,28,198,78,72,3.19231,4.04167,0,2,0,467,
-expected_impact_archery_standard_baseline,3536,/private/tmp/eagle0-standard-baselines/configs/3536_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,12,120,57,36,3.40351,2.88889,4,0,2528,0,
-expected_impact_archery_standard_baseline,3701,/private/tmp/eagle0-standard-baselines/configs/3701_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,30,155,62,70,3.5,3.24286,0,4,0,1348,
-expected_impact_archery_standard_baseline,4298,/private/tmp/eagle0-standard-baselines/configs/4298_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,3,38,13,11,3.07692,2.90909,0,6,0,1977,
-expected_impact_archery_standard_baseline,4576,/private/tmp/eagle0-standard-baselines/configs/4576_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,4,35,17,7,3,3.28571,0,3,0,1986,
-expected_impact_archery_standard_baseline,6014,/private/tmp/eagle0-standard-baselines/configs/6014_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,6,39,20,7,3.15,3.42857,0,3,0,2252,
-expected_impact_archery_standard_baseline,6248,/private/tmp/eagle0-standard-baselines/configs/6248_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,25,153,75,46,3.65333,2.8913,2,0,676,0,
-expected_impact_archery_standard_baseline,6408,/private/tmp/eagle0-standard-baselines/configs/6408_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,11,107,68,18,3.5,3.33333,5,0,4020,0,
-expected_impact_archery_standard_baseline,6705,/private/tmp/eagle0-standard-baselines/configs/6705_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,31,205,87,86,3.01149,3.74419,0,2,0,535,
-expected_impact_archery_standard_baseline,7212,/private/tmp/eagle0-standard-baselines/configs/7212_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,8,39,18,11,3.38889,3.63636,0,2,0,1273,
-expected_impact_archery_standard_baseline,8217,/private/tmp/eagle0-standard-baselines/configs/8217_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,22,164,79,49,3.35443,3.08163,3,0,2167,0,
-expected_impact_archery_standard_baseline,8394,/private/tmp/eagle0-standard-baselines/configs/8394_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,9,81,48,17,3.25,2.64706,4,0,3186,0,
-expected_impact_archery_standard_baseline,8628,/private/tmp/eagle0-standard-baselines/configs/8628_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,5,41,27,3,2.96296,2,4,0,2760,0,
-expected_impact_archery_standard_baseline,9975,/private/tmp/eagle0-standard-baselines/configs/9975_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,4,70,23,27,2.65217,2.40741,0,10,0,7840,
-expected_impact_archery_standard_baseline,10229,/private/tmp/eagle0-standard-baselines/configs/10229_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,23,104,57,4,3.77193,3,3,0,1816,0,
-expected_impact_archery_standard_baseline,10631,/private/tmp/eagle0-standard-baselines/configs/10631_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,1,16,7,1,3,3,5,1,4000,0,
-expected_impact_archery_standard_baseline,11011,/private/tmp/eagle0-standard-baselines/configs/11011_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,5,51,22,18,3.5,3,0,5,0,3658,
-expected_impact_archery_standard_baseline,11012,/private/tmp/eagle0-standard-baselines/configs/11012_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,30,146,47,67,3.40426,3.62687,0,3,0,1347,
-expected_impact_archery_standard_baseline,11112,/private/tmp/eagle0-standard-baselines/configs/11112_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,3,31,7,8,3.42857,2,0,11,0,7595,
-expected_impact_archery_standard_baseline,11297,/private/tmp/eagle0-standard-baselines/configs/11297_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,1,13,5,1,3,3,3,1,2400,0,
-expected_impact_archery_standard_baseline,11469,/private/tmp/eagle0-standard-baselines/configs/11469_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,8,58,30,12,3.16667,3.16667,0,2,0,1489,
-expected_impact_archery_standard_baseline,11470,/private/tmp/eagle0-standard-baselines/configs/11470_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,17,99,65,22,3.50769,2.68182,4,0,3425,0,
-expected_impact_archery_standard_baseline,11565,/private/tmp/eagle0-standard-baselines/configs/11565_standard.json,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,103,30,34,4.26667,3.91176,1,4,78,3600,
-expected_impact_archery_standard_baseline,11662,/private/tmp/eagle0-standard-baselines/configs/11662_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,8,120,85,17,2.85882,3.23529,9,0,6624,0,
-expected_impact_archery_standard_baseline,11754,/private/tmp/eagle0-standard-baselines/configs/11754_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,30,163,72,76,3.31944,3.28947,0,3,0,1000,
-expected_impact_archery_standard_baseline,11941,/private/tmp/eagle0-standard-baselines/configs/11941_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,6,64,44,7,3.25,3,6,0,4777,0,
-expected_impact_archery_standard_baseline,11942,/private/tmp/eagle0-standard-baselines/configs/11942_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,5,91,59,16,2.64407,3.375,12,0,8572,0,
-expected_impact_archery_standard_baseline,12136,/private/tmp/eagle0-standard-baselines/configs/12136_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,21,141,79,34,3.70886,4.52941,3,0,1581,0,
-expected_impact_archery_standard_baseline,12137,/private/tmp/eagle0-standard-baselines/configs/12137_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,8,50,19,17,3.21053,3.29412,0,3,0,1892,
-expected_impact_archery_standard_baseline,12138,/private/tmp/eagle0-standard-baselines/configs/12138_standard.json,attacker,last_player_standing,Attacker won by eliminating all defenders,14,102,69,16,3.43478,2.9375,3,0,2512,0,
-expected_impact_archery_standard_baseline,12765,/private/tmp/eagle0-standard-baselines/configs/12765_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,11,46,26,11,3.61538,3.09091,0,2,0,1763,
-expected_impact_archery_standard_baseline,13752,/private/tmp/eagle0-standard-baselines/configs/13752_standard.json,defender,last_player_standing,Defender won by eliminating all attackers,7,38,16,10,3.5,3.3,0,2,0,932,
@@ -1,40 +0,0 @@
-# STANDARD Scoring All-40 Baseline
-
-This baseline is the current comparison baseline for STANDARD scoring experiments. It was
-regenerated after fixing iterative-deepening timeout score propagation.
-Generated battle configs preserve archery/start-fire capability, per-unit flee capability,
-explicit zero-size battalions, no-profession heroes, forced commands, seed `1`, and the
-31-round defender timeout.
-
-Raw rows are recorded in `standard_scoring_all40_post_bugfix_baseline.csv`.
-
-## Summary
-
-| Metric | Count |
-| --- | ---: |
-| Real battles | 40 |
-| Valid simulated battles | 40 |
-| Invalid / no recoverable troops | 0 |
-| Attacker wins | 18 |
-| Defender wins | 22 |
-| Last-player-standing endings | 39 |
-| Max-round defender endings | 1 |
-
-## Closest Battles By Survivor Margin
-
-Survivor margin is `attacker_surviving_troops - defender_surviving_troops`.
-
-| Action seq | Winner | End reason | Attacker troops | Defender troops | Margin |
-| ---: | --- | --- | ---: | ---: | ---: |
-| 2584 | defender | last_player_standing | 0 | 168 | -168 |
-| 2654 | defender | last_player_standing | 0 | 467 | -467 |
-| 6705 | defender | last_player_standing | 0 | 535 | -535 |
-| 307 | defender | last_player_standing | 0 | 648 | -648 |
-| 6248 | attacker | last_player_standing | 676 | 0 | +676 |
-| 13752 | defender | last_player_standing | 0 | 932 | -932 |
-| 1472 | defender | last_player_standing | 0 | 944 | -944 |
-| 11754 | defender | last_player_standing | 0 | 1000 | -1000 |
-| 7212 | defender | last_player_standing | 0 | 1273 | -1273 |
-| 399 | defender | last_player_standing | 0 | 1315 | -1315 |
-| 11012 | defender | last_player_standing | 0 | 1347 | -1347 |
-| 3701 | defender | last_player_standing | 0 | 1348 | -1348 |
@@ -1,41 +0,0 @@
-action_seq,attacker_factions,attacker_player,defender_factions,defender_player,ai_vs_ai,real_winner,sim_winner,winner_match,real_round,sim_round,real_attacker_troops,real_defender_troops,sim_attacker_troops,sim_defender_troops
-306,1,AI,2,AI,true,attacker,attacker,true,0,19,1454,0,1388,0
-307,2,AI,1,AI,true,defender,defender,true,0,16,0,699,0,648
-399,2,AI,1,AI,true,defender,defender,true,0,6,0,1310,0,1315
-631,1,AI,2,AI,true,defender,defender,true,0,5,0,3766,0,3654
-1326,9,AI,6,AI,true,attacker,attacker,true,0,11,1715,0,1701,0
-1472,8,AI,3,human,false,defender,defender,true,0,31,0,1593,0,944
-2437,2,AI,1,AI,true,attacker,attacker,true,0,16,3146,0,3247,0
-2584,1,AI,2,AI,true,defender,defender,true,0,31,0,2028,0,168
-2654,2,AI,1,AI,true,defender,defender,true,0,28,0,335,0,467
-3536,1,AI,2,AI,true,attacker,attacker,true,0,12,2172,0,2528,0
-3701,8,AI,2,AI,true,defender,defender,true,0,30,0,2819,0,1348
-4298,2,AI,1,AI,true,defender,defender,true,0,3,0,3514,0,1977
-4576,2,AI,1,AI,true,defender,defender,true,0,4,0,2028,0,1986
-6014,2,AI,8,AI,true,defender,defender,true,0,6,0,1942,0,2252
-6248,7,AI,9,AI,true,attacker,attacker,true,0,25,520,0,676,0
-6408,4,human,7,AI,false,attacker,attacker,true,0,11,4156,0,4020,0
-6705,1,AI,2,AI,true,defender,defender,true,0,31,0,3316,0,535
-7212,2,AI,8,AI,true,defender,defender,true,0,8,0,1253,0,1273
-8217,8,AI,2,AI,true,attacker,attacker,true,0,22,920,0,2167,0
-8394,2,AI,8,AI,true,attacker,attacker,true,0,9,2449,0,3186,0
-8628,3,human,8,AI,false,attacker,attacker,true,0,5,2874,0,2760,0
-9975,2,AI,1,AI,true,defender,defender,true,0,4,0,7420,0,7840
-10229,5,AI,7,AI,true,attacker,attacker,true,0,23,2135,0,1816,0
-10631,1,AI,2,AI,true,attacker,attacker,true,0,1,4000,0,4000,0
-11011,2,AI,1,AI,true,defender,defender,true,0,5,0,3532,0,3658
-11012,9,AI,6,AI,true,defender,defender,true,0,30,0,846,0,1347
-11112,5,AI,4,human,false,defender,defender,true,0,3,0,7595,0,7595
-11297,1,AI,2,AI,true,attacker,attacker,true,0,1,2400,0,2400,0
-11469,5,AI,9,AI,true,defender,defender,true,0,8,0,787,0,1489
-11470,7,AI,5,AI,true,attacker,attacker,true,0,17,3220,0,3425,0
-11565,5,AI,7,AI,true,defender,defender,true,0,31,190,3420,78,3600
-11662,4,human,9,AI,false,attacker,attacker,true,0,8,7326,0,6624,0
-11754,7,AI,4,human,false,attacker,defender,false,0,30,2119,0,0,1000
-11941,3,human,5,AI,false,attacker,attacker,true,0,6,4901,0,4777,0
-11942,4,human,7,AI,false,attacker,attacker,true,0,5,8840,0,8572,0
-12136,9,AI,7,AI,true,attacker,attacker,true,0,21,1829,0,1581,0
-12137,2,AI,1,AI,true,defender,defender,true,0,8,0,1894,0,1892
-12138,3,human,5,AI,false,attacker,attacker,true,0,14,2960,0,2512,0
-12765,6,AI,9,AI,true,defender,defender,true,0,11,0,1773,0,1763
-13752,2,AI,1,AI,true,defender,defender,true,0,7,0,943,0,932
@@ -1,56 +0,0 @@
-# All-40 Real vs STANDARD Simulation
-
-This table compares the saved-game Shardok winner side with a fresh
-STANDARD-vs-STANDARD simulator run from the extracted battle setup.
-Faction `3` and faction `4` are marked as human players; all other
-factions are marked as AI players.
-
- Battles checked: 40
- Known real outcomes: 40
- Winner-side matches on known outcomes: 39/40
- Real saved-game winners: 19 attacker, 21 defender, 0 unknown
- Sim baseline winners: 18 attacker, 22 defender
- Battles involving a human side: 9
-
-| Action seq | Attacker | Defender | Real winner | Sim winner | Match | Real troops | Sim troops |
-| ---: | --- | --- | --- | --- | --- | ---: | ---: |
-| 306 | 1 (AI) | 2 (AI) | attacker | attacker | yes | 1454 vs 0 | 1388 vs 0 |
-| 307 | 2 (AI) | 1 (AI) | defender | defender | yes | 0 vs 699 | 0 vs 648 |
-| 399 | 2 (AI) | 1 (AI) | defender | defender | yes | 0 vs 1310 | 0 vs 1315 |
-| 631 | 1 (AI) | 2 (AI) | defender | defender | yes | 0 vs 3766 | 0 vs 3654 |
-| 1326 | 9 (AI) | 6 (AI) | attacker | attacker | yes | 1715 vs 0 | 1701 vs 0 |
-| 1472 | 8 (AI) | 3 (human) | defender | defender | yes | 0 vs 1593 | 0 vs 944 |
-| 2437 | 2 (AI) | 1 (AI) | attacker | attacker | yes | 3146 vs 0 | 3247 vs 0 |
-| 2584 | 1 (AI) | 2 (AI) | defender | defender | yes | 0 vs 2028 | 0 vs 168 |
-| 2654 | 2 (AI) | 1 (AI) | defender | defender | yes | 0 vs 335 | 0 vs 467 |
-| 3536 | 1 (AI) | 2 (AI) | attacker | attacker | yes | 2172 vs 0 | 2528 vs 0 |
-| 3701 | 8 (AI) | 2 (AI) | defender | defender | yes | 0 vs 2819 | 0 vs 1348 |
-| 4298 | 2 (AI) | 1 (AI) | defender | defender | yes | 0 vs 3514 | 0 vs 1977 |
-| 4576 | 2 (AI) | 1 (AI) | defender | defender | yes | 0 vs 2028 | 0 vs 1986 |
-| 6014 | 2 (AI) | 8 (AI) | defender | defender | yes | 0 vs 1942 | 0 vs 2252 |
-| 6248 | 7 (AI) | 9 (AI) | attacker | attacker | yes | 520 vs 0 | 676 vs 0 |
-| 6408 | 4 (human) | 7 (AI) | attacker | attacker | yes | 4156 vs 0 | 4020 vs 0 |
-| 6705 | 1 (AI) | 2 (AI) | defender | defender | yes | 0 vs 3316 | 0 vs 535 |
-| 7212 | 2 (AI) | 8 (AI) | defender | defender | yes | 0 vs 1253 | 0 vs 1273 |
-| 8217 | 8 (AI) | 2 (AI) | attacker | attacker | yes | 920 vs 0 | 2167 vs 0 |
-| 8394 | 2 (AI) | 8 (AI) | attacker | attacker | yes | 2449 vs 0 | 3186 vs 0 |
-| 8628 | 3 (human) | 8 (AI) | attacker | attacker | yes | 2874 vs 0 | 2760 vs 0 |
-| 9975 | 2 (AI) | 1 (AI) | defender | defender | yes | 0 vs 7420 | 0 vs 7840 |
-| 10229 | 5 (AI) | 7 (AI) | attacker | attacker | yes | 2135 vs 0 | 1816 vs 0 |
-| 10631 | 1 (AI) | 2 (AI) | attacker | attacker | yes | 4000 vs 0 | 4000 vs 0 |
-| 11011 | 2 (AI) | 1 (AI) | defender | defender | yes | 0 vs 3532 | 0 vs 3658 |
-| 11012 | 9 (AI) | 6 (AI) | defender | defender | yes | 0 vs 846 | 0 vs 1347 |
-| 11112 | 5 (AI) | 4 (human) | defender | defender | yes | 0 vs 7595 | 0 vs 7595 |
-| 11297 | 1 (AI) | 2 (AI) | attacker | attacker | yes | 2400 vs 0 | 2400 vs 0 |
-| 11469 | 5 (AI) | 9 (AI) | defender | defender | yes | 0 vs 787 | 0 vs 1489 |
-| 11470 | 7 (AI) | 5 (AI) | attacker | attacker | yes | 3220 vs 0 | 3425 vs 0 |
-| 11565 | 5 (AI) | 7 (AI) | defender | defender | yes | 190 vs 3420 | 78 vs 3600 |
-| 11662 | 4 (human) | 9 (AI) | attacker | attacker | yes | 7326 vs 0 | 6624 vs 0 |
-| 11754 | 7 (AI) | 4 (human) | attacker | defender | no | 2119 vs 0 | 0 vs 1000 |
-| 11941 | 3 (human) | 5 (AI) | attacker | attacker | yes | 4901 vs 0 | 4777 vs 0 |
-| 11942 | 4 (human) | 7 (AI) | attacker | attacker | yes | 8840 vs 0 | 8572 vs 0 |
-| 12136 | 9 (AI) | 7 (AI) | attacker | attacker | yes | 1829 vs 0 | 1581 vs 0 |
-| 12137 | 2 (AI) | 1 (AI) | defender | defender | yes | 0 vs 1894 | 0 vs 1892 |
-| 12138 | 3 (human) | 5 (AI) | attacker | attacker | yes | 2960 vs 0 | 2512 vs 0 |
-| 12765 | 6 (AI) | 9 (AI) | defender | defender | yes | 0 vs 1773 | 0 vs 1763 |
-| 13752 | 2 (AI) | 1 (AI) | defender | defender | yes | 0 vs 943 | 0 vs 932 |
@@ -1,70 +0,0 @@
-# ID Scoring Experiments, Seeds 1-4
-
-These results use `id_standard_seed1.json` as the base generated battle shape:
-
- Map: `Alah`
- Max rounds: 10
- Setup: AI placement, forced single-option commands enabled
- Attacker: iterative deepening, six longbowmen with professions 1-6
- Defender: iterative deepening, six light infantry units with no attached heroes
-
-The saved game zip at the repository root contains real Shardok battles, but the available
-`shardok_state` rows are latest-state snapshots and the sampled states are terminal. They are
-useful for saved-state ingestion checks, not for live AI decision win-rate measurement.
-
-## Baseline
-
-`STANDARD` attacker vs `STANDARD` defender:
-
- Attacker wins: 0 / 4
- Defender wins: 4 / 4
- Attacker win rate: 0%
- Average defender surviving units: 6.00
- Average defender surviving troops: 3510.25
-
-Raw rows are recorded in `id_standard_baseline_seed1_4.csv`.
-
-## Candidate Results
-
-| Attacker scoring | Defender scoring | Attacker wins | Defender wins | Avg defender units | Avg defender troops | Notes |
-| --- | --- | ---: | ---: | ---: | ---: | --- |
-| `NORMALIZED` | `STANDARD` | 0 | 4 | 5.50 | 3224.00 | Better casualties than baseline, no win-rate movement. |
-| `MCTS_OPTIMIZED` | `STANDARD` | 0 | 4 | 5.50 | 3503.50 | Similar troop outcome to baseline. |
-| `STANDARD` with attacker unit decay removed | `STANDARD` | 0 | 4 | 6.00 | 3204.25 | Best casualty signal in this small sample, no win-rate movement. |
-
-The attacker unit-decay experiment changes `StandardAIScoreCalculator::CombineAttackerScores` to
-use a constant unit multiplier, matching the defender hold-castle scoring behavior. The rationale is
-the same as the existing defender-side comment: decaying unit value late in the battle can make
-tactical actions look worse than waiting when a side is behind on unit value.
-
-## Candidate Raw Rows
-
-### `NORMALIZED` Attacker vs `STANDARD` Defender
-
-```csv
-seed,attacker_scoring,defender_scoring,winner,end_reason,total_rounds,total_commands,attacker_ai_commands,defender_ai_commands,attacker_avg_depth,defender_avg_depth,attacker_surviving_units,defender_surviving_units,attacker_surviving_troops,defender_surviving_troops
-1,NORMALIZED,STANDARD,defender,last_player_standing,9,121,60,43,2.53333,3.44186,0,5,0,2665
-2,NORMALIZED,STANDARD,defender,last_player_standing,10,129,64,47,2.50000,3.46809,0,6,0,3102
-3,NORMALIZED,STANDARD,defender,last_player_standing,9,116,59,41,2.59322,3.31707,0,6,0,3752
-4,NORMALIZED,STANDARD,defender,last_player_standing,9,114,63,34,2.47619,3.47059,0,5,0,3377
-```
-
-### `MCTS_OPTIMIZED` Attacker vs `STANDARD` Defender
-
-```csv
-seed,attacker_scoring,defender_scoring,winner,end_reason,total_rounds,total_commands,attacker_ai_commands,defender_ai_commands,attacker_avg_depth,defender_avg_depth,attacker_surviving_units,defender_surviving_units,attacker_surviving_troops,defender_surviving_troops
-1,MCTS_OPTIMIZED,STANDARD,defender,last_player_standing,9,129,64,46,2.48438,3.54348,0,6,0,3203
-2,MCTS_OPTIMIZED,STANDARD,defender,last_player_standing,9,119,62,40,2.51613,3.22500,0,6,0,3559
-3,MCTS_OPTIMIZED,STANDARD,defender,last_player_standing,9,115,62,35,2.67742,3.37143,0,5,0,3969
-4,MCTS_OPTIMIZED,STANDARD,defender,last_player_standing,9,114,63,34,2.41270,3.41176,0,5,0,3283
-```
-
-### `STANDARD` Attacker With Unit Decay Removed vs `STANDARD` Defender
-
-```csv
-seed,attacker_scoring,defender_scoring,winner,end_reason,total_rounds,total_commands,attacker_ai_commands,defender_ai_commands,attacker_avg_depth,defender_avg_depth,attacker_surviving_units,defender_surviving_units,attacker_surviving_troops,defender_surviving_troops
-1,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,defender,last_player_standing,10,135,65,51,2.53846,3.45098,0,6,0,2707
-2,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,defender,last_player_standing,9,130,64,48,2.48438,3.58333,0,6,0,3149
-3,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,defender,last_player_standing,9,129,63,45,2.57143,3.44444,0,6,0,3401
-4,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,defender,last_player_standing,9,129,62,46,2.66129,3.41304,0,6,0,3560
-```
@@ -1,5 +0,0 @@
-seed,attacker_scoring,defender_scoring,winner,end_reason,total_rounds,total_commands,attacker_ai_commands,defender_ai_commands,attacker_avg_depth,defender_avg_depth,attacker_surviving_units,defender_surviving_units,attacker_surviving_troops,defender_surviving_troops
-1,STANDARD,STANDARD,defender,last_player_standing,9,130,65,47,2.52308,3.61702,0,6,0,3376
-2,STANDARD,STANDARD,defender,last_player_standing,9,130,63,47,2.58730,3.29787,0,6,0,3704
-3,STANDARD,STANDARD,defender,last_player_standing,9,129,63,45,2.58730,3.44444,0,6,0,3401
-4,STANDARD,STANDARD,defender,last_player_standing,9,129,62,46,2.70968,3.39130,0,6,0,3560
@@ -1,32 +0,0 @@
-{
-  "map_name": "Alah",
-  "month": 4,
-  "max_rounds": 10,
-  "random_seed": 1,
-  "use_configured_placements": false,
-  "apply_forced_commands": true,
-  "attacker": {
-    "ai_algorithm": "ITERATIVE_DEEPENING",
-    "scoring_calculator": "STANDARD",
-    "units": [
-      { "profession": 1, "battalion_type_id": 4, "starting_position_index": 0, "battalion": { "size": 1000 } },
-      { "profession": 2, "battalion_type_id": 4, "starting_position_index": 0, "battalion": { "size": 1000 } },
-      { "profession": 3, "battalion_type_id": 4, "starting_position_index": 0, "battalion": { "size": 1000 } },
-      { "profession": 4, "battalion_type_id": 4, "starting_position_index": 0, "battalion": { "size": 1000 } },
-      { "profession": 5, "battalion_type_id": 4, "starting_position_index": 0, "battalion": { "size": 1000 } },
-      { "profession": 6, "battalion_type_id": 4, "starting_position_index": 0, "battalion": { "size": 1000 } }
-    ]
-  },
-  "defender": {
-    "ai_algorithm": "ITERATIVE_DEEPENING",
-    "scoring_calculator": "STANDARD",
-    "units": [
-      { "profession": 14, "battalion_type_id": 0, "starting_position_index": -1, "battalion": { "size": 1000 } },
-      { "profession": 14, "battalion_type_id": 0, "starting_position_index": -1, "battalion": { "size": 1000 } },
-      { "profession": 14, "battalion_type_id": 0, "starting_position_index": -1, "battalion": { "size": 1000 } },
-      { "profession": 14, "battalion_type_id": 0, "starting_position_index": -1, "battalion": { "size": 1000 } },
-      { "profession": 14, "battalion_type_id": 0, "starting_position_index": -1, "battalion": { "size": 1000 } },
-      { "profession": 14, "battalion_type_id": 0, "starting_position_index": -1, "battalion": { "size": 1000 } }
-    ]
-  }
-}
@@ -1,41 +0,0 @@
-variant,attacker_scoring,defender_scoring,shardok_game_id,action_seq,attacker_units,defender_units,attacker_initial_troops,defender_initial_troops,attacker_troop_ratio,winner,end_reason,result,total_rounds,total_commands,attacker_ai_commands,defender_ai_commands,attacker_avg_depth,defender_avg_depth,attacker_surviving_units,defender_surviving_units,attacker_surviving_troops,defender_surviving_troops,elapsed_seconds
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_1_b3e7d8ab,306,2,3,1600.0,2600.0,0.6153846153846154,defender,last_player_standing,Defender won by eliminating all attackers,31,192,67,85,3.61194,3.61176,0,1,0,685,18.768
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_2_bf499d3c,307,2,1,1600.0,1000.0,1.6,defender,last_player_standing,Defender won by eliminating all attackers,20,82,32,20,3.09375,2.3,0,1,0,324,5.722
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_3_cd9141a,399,1,3,450.0,1454.0,0.30949105914718017,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,188,45,106,2,3.89623,1,3,450,2454,30.201
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_4_6e7b9b,631,2,5,1310.0,4200.0,0.3119047619047619,defender,last_player_standing,Defender won by eliminating all attackers,8,70,20,33,3.55,3.90909,0,4,0,2991,9.564
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_5_8a57255a,1326,2,4,1800.0,0.0,,defender,last_player_standing,Defender won by eliminating all attackers,9,61,18,27,2.77778,3.18519,0,5,0,4037,10.772
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_6_ab8632ba,1472,2,2,1800.0,1600.0,1.125,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,198,69,84,3.13043,3.5,1,2,57,378,23.999
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_7_5cf74e0,2437,5,6,3766.0,3787.0,0.9944547134935305,defender,last_player_standing,Defender won by eliminating all attackers,18,267,82,158,3.30488,3.42405,0,5,0,2948,73.020
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_8_41ce0ebd,2584,2,7,1600.0,3668.0,0.4362050163576881,defender,last_player_standing,Defender won by eliminating all attackers,6,78,10,52,3.2,3.53846,0,7,0,3832,21.143
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_9_eb43d421,2654,3,2,2599.0,726.0,3.5798898071625342,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,317,186,93,2,4.12903,3,2,2599,726,33.775
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_a_950de1c1,3536,5,10,3858.0,2028.0,1.9023668639053255,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,428,138,247,2.52174,3.39271,1,6,375,4511,115.585
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_b_5d1d76e0,3701,2,6,1800.0,4240.0,0.42452830188679247,defender,last_player_standing,Defender won by eliminating all attackers,31,255,40,179,2.825,3.45251,0,4,0,2358,56.109
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_c_7bdb54dd,4298,4,6,3123.0,2172.0,1.4378453038674033,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,599,247,309,2,3.69579,4,6,2914,4085,128.531
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_d_689ee2fa,4576,4,3,3300.0,2135.0,1.5456674473067915,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,415,210,148,2,3.33108,4,3,629,1771,41.509
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_e_89c90dc6,6014,2,3,1512.0,2600.0,0.5815384615384616,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,214,97,79,2,3.29114,2,3,1512,2600,27.933
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_f_3a9d9136,6248,2,2,1800.0,1600.0,1.125,attacker,last_player_standing,Attacker won by eliminating all defenders,12,86,32,27,3.25,3.37037,2,0,1455,0,7.670
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_10_c972d3e1,6408,5,2,4169.0,1223.0,3.4088307440719543,attacker,last_player_standing,Attacker won by eliminating all defenders,15,145,96,24,3.38542,2.54167,5,0,3832,0,30.677
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_11_e0611233,6705,5,5,3510.0,4400.0,0.7977272727272727,attacker,last_player_standing,Attacker won by eliminating all defenders,28,219,101,81,2.9802,3.48148,3,0,2150,0,40.168
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_12_84a36154,7212,2,2,1156.0,1800.0,0.6422222222222222,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,254,62,124,2,3.45161,2,2,1156,1800,15.591
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_13_2e98b1cf,8217,3,2,2270.0,1136.0,1.9982394366197183,attacker,last_player_standing,Attacker won by eliminating all defenders,24,185,90,53,3.46667,3.37736,3,0,1944,0,21.446
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_14_b5570863,8394,4,2,3600.0,1212.0,2.9702970297029703,attacker,last_player_standing,Attacker won by eliminating all defenders,20,130,74,26,3.41892,2.84615,4,0,2552,0,25.721
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_15_28a14d7d,8628,4,1,3011.0,800.0,3.76375,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,255,186,31,2,3,4,1,3011,800,5.159
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_16_3d6ec341,9975,5,11,4049.0,9110.0,0.44445664105378707,defender,last_player_standing,Defender won by eliminating all attackers,5,115,26,65,3.03846,2.98462,0,11,0,7033,33.494
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_17_3ba689e3,10229,2,1,1800.0,1000.0,1.8,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,160,81,13,3.85185,2.53846,1,1,56,727,15.586
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_18_9e463db3,10631,5,2,4000.0,29.0,137.93103448275863,attacker,last_player_standing,Attacker won by eliminating all defenders,10,104,55,27,2.98182,2.62963,5,0,3955,0,17.981
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_19_239725d3,11011,3,5,2092.0,4000.0,0.523,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,274,80,151,2.3625,3.75497,1,4,120,2749,65.354
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_1a_16e1ebf,11012,2,2,1600.0,1800.0,0.8888888888888888,defender,last_player_standing,Defender won by eliminating all attackers,21,112,30,48,2.9,3.3125,0,2,0,1328,12.932
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_1b_814a8b05,11112,2,11,1800.0,7595.0,0.2369980250164582,defender,last_player_standing,Defender won by eliminating all attackers,5,89,10,60,3.4,3.05,0,11,0,9248,26.113
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_1c_529ba6e5,11297,3,2,2400.0,0.0,,attacker,last_player_standing,Attacker won by eliminating all defenders,14,110,46,39,2.95652,3.4359,3,0,1569,0,20.096
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_1d_49993012,11469,3,2,2383.0,1800.0,1.323888888888889,attacker,last_player_standing,Attacker won by eliminating all defenders,27,159,73,34,3.52055,3.05882,3,0,2160,0,19.694
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_1e_fd8c60a8,11470,4,2,3600.0,1800.0,2.0,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,356,177,112,2.31073,3.47321,3,2,2267,952,18.046
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_1f_53309484,11565,1,4,334.0,3600.0,0.09277777777777778,defender,last_player_standing,Defender won by eliminating all attackers,1,8,1,0,3,0,0,4,0,3600,0.555
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_20_6d9f21cd,11662,9,3,7595.0,2600.0,2.921153846153846,attacker,last_player_standing,Attacker won by eliminating all defenders,11,174,111,41,3.1982,3.36585,9,0,7070,0,47.504
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_21_c536f279,11754,3,3,2470.0,2400.0,1.0291666666666666,defender,last_player_standing,Defender won by eliminating all attackers,7,55,20,22,3.65,3.54545,0,3,0,1624,10.181
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_22_a0f39b09,11941,6,2,4996.0,1800.0,2.7755555555555556,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,290,186,69,3.46237,3.52174,1,2,165,299,60.206
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_23_865a7a0e,11942,12,3,8916.0,2119.0,4.207645115620576,attacker,last_player_standing,Attacker won by eliminating all defenders,12,232,164,47,2.94512,3.59574,12,0,7940,0,77.700
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_24_f0d61a5a,12136,3,2,2089.0,1800.0,1.1605555555555556,attacker,last_player_standing,Attacker won by eliminating all defenders,15,98,50,25,3.18,3.12,2,0,1384,0,14.063
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_25_a19c4d87,12137,2,3,1524.0,2400.0,0.635,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,148,67,43,2.34328,2.55814,1,3,214,2282,19.123
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_26_d897bec7,12138,4,2,3301.0,1400.0,2.3578571428571427,defender,max_rounds_reached,Defender won by surviving maximum rounds,31,422,290,93,2,3.33333,4,2,3301,1400,6.678
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_27_c008989b,12765,2,2,1800.0,1800.0,1.0,defender,last_player_standing,Defender won by eliminating all attackers,30,108,40,30,3.5,2.86667,0,2,0,1678,9.069
-no_attacker_unit_decay,STANDARD_NO_ATTACKER_UNIT_DECAY,STANDARD,389ffb1901903118_28_b33e2868,13752,2,2,1798.0,1600.0,1.12375,defender,last_player_standing,Defender won by eliminating all attackers,30,189,39,123,3.15385,3.98374,0,2,0,1114,26.389
--- a/Show More
+++ b/Show More