loop-benchmarking

commit 669aa68861617a446e5ae56aea0a462995d183f0
parent 71f0c4b7931a2d58ae370587a6a0f62ba352b716
Author: Brian Graham <brian@buildingbetterteams.de>
Date:   Sat, 11 Apr 2026 07:30:25 +0200

V2: game_over_display test passes on overlay OR restart presence

Previous check required BOTH a modal and a restart button. Now accepts
either signal because different games show game-over UI differently
(some have full modal, some just show a restart button overlay).

Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>

Diffstat:
M
tasks/tetris/eval/gameplay-bot-v2/bot.ts
 | 
30
++++++++++++++++++------------

1 file changed, 18 insertions(+), 12 deletions(-)
diff --git a/tasks/tetris/eval/gameplay-bot-v2/bot.ts b/tasks/tetris/eval/gameplay-bot-v2/bot.ts
@@ -1873,30 +1873,36 @@ function deriveTestResults(
   }
 
   // 22. game_over_display
-  // Prefer the Phase 6 capture (taken immediately after game over was
-  // triggered) because Phase 8's competitive-play check happens after a fresh
-  // page reload and rarely sees a real game-over screen. Fall back to the
-  // Phase 8 reading only if Phase 6 never reached game over.
+  // Verifies the game SHOWS a game-over UI after the internal game over
+  // state is reached. Two structural signals (no language matching):
+  //   1. A modal or overlay element appeared (position fixed/absolute,
+  //      covering >15% viewport, visible). Captured in Phase 6 after
+  //      game over was triggered via grid stacking.
+  //   2. A clickable element (restart) is visible inside that overlay.
+  // The test passes if EITHER a modal is present OR a restart option is
+  // visible, because different games show different UI styles.
   if (!phaseState.gameplayWorks || !competitivePlay) {
     results.push(skipResult("game_over_display", "competitive play phase did not run"));
   } else if (!competitivePlay.game_over_reached && !session.gameOverDetected) {
     results.push(skipResult("game_over_display", "game over not reached during play"));
   } else {
     const usePhase6 = session.gameOverDetected;
-    const text = usePhase6
-      ? session.gameOverText ?? null
-      : competitivePlay.game_over_text_found;
+    const modalFound = usePhase6
+      ? (session.gameOverText !== null && session.gameOverText !== undefined)
+      : (competitivePlay.game_over_text_found !== null);
     const hasRestart = usePhase6
       ? session.gameOverRestartAvailable === true
       : competitivePlay.restart_available;
-    const hasText = text !== null && text !== undefined;
     const source = usePhase6 ? "phase6" : "phase8";
+    const pass = modalFound || hasRestart;
+    const details: string[] = [];
+    if (modalFound) details.push("overlay detected");
+    if (hasRestart) details.push("restart clickable present");
+    if (!pass) details.push("no overlay or restart UI found");
     results.push({
       name: "game_over_display",
-      pass: hasText && hasRestart,
-      detail: hasText && hasRestart
-        ? `game over display: "${text}", restart available (${source})`
-        : `missing: ${!hasText ? "game over text" : ""}${!hasText && !hasRestart ? " and " : ""}${!hasRestart ? "restart option" : ""} (${source})`,
+      pass,
+      detail: `${details.join(", ")} (${source})`,
     });
   }