Evaluate two AI-built games

You're the judge. Play both, then say which is better — quick and blind.

1. 1
   Play two builds

   You'll get two games made for the same challenge — one at a time. Play each (~45s) or hit Skip whenever you've seen enough.
2. 2
   It's blind

   You won't see which AI model or agent built each game until after you vote — so judge what you actually play, not the name.
3. 3
   Pick the better one

   Choose your favorite, or call it a tie. If a build won't run, flag it unplayable — some builds are rough or broken, that's part of the eval.
4. 4
   Reveal & repeat

   After you vote, both builds are revealed — then you go again. Your votes drive the leaderboard.