Context:
Resolution: Resolves to the lab responsible for the AI model that is ranked highest on the Kaggle Game Arena leaderboard for Poker at EOD February 4.
Kaggle hasn't released a ton of details about this, so trade with caution and I may need to clarify things if issues arise.
Update 2026-02-02 (PST) (AI summary of creator comment): Fallback resolution method clarified:
If the leaderboard has been filled in by EOD February 4, resolution will be based on the leaderboard
If the leaderboard has NOT been filled in by EOD February 4, resolution will be based on the results of the poker tournament bracket
Creator expects these two methods will yield the same result
🏅 Top traders
| # | Trader | Total profit |
|---|---|---|
| 1 | Ṁ564 | |
| 2 | Ṁ179 | |
| 3 | Ṁ124 | |
| 4 | Ṁ118 | |
| 5 | Ṁ112 |
https://www.kaggle.com/benchmarks/kaggle/poker-heads-up/versions/1/leaderboard they posted the leaderboard, presumably based on the tournament gameplay. openai wins
Kaggle hasn't released a ton of details about this, so trade with caution and I may need to clarify things if issues arise.
Ok they released some more details: https://x.com/kaggle/status/2018356777167826997
It looks like Poker is initially a heads up bracket: https://www.kaggle.com/benchmarks/kaggle/poker-heads-up/versions/1/tournament. The first round is done, and there are 4 models in the semi-finals (including two OpenAI models).
Technically, there's a separate tab for the leaderboard:
Leaderboard coming soon
Hang tight! The leaderboard will be ready for you some time after the tournament wraps up.
Until then, keep up with all the exciting live matches on our game bracket!
Here's what I said in the description:
Resolution: Resolves to the lab responsible for the AI model that is ranked highest on the Kaggle Game Arena leaderboard for Poker at EOD February 4.
The goal is to evaluate the results as of EOD Feb 4. Given that, if the leaderboard hasn't been filled in by then, I'll resolve based on the results of the poker tourney. If the leaderboard HAS been filled in by then, I'll use the leaderboard. My expectation is that these will be the same, as the leaderboard is going to be based (at least initially?) on the heads up tournament? But I could be wrong.
