AI beats Stratego's greatest player, ending one of the last human strongholds in board games
Ataraxos beat Stratego champion Niemeijer using far less compute than DeepMind's DeepNash.
University researchers' Ataraxos beat Stratego great Niemeijer with an 85% effective win rate, counting draws as half wins, over a three-week series. It trained one week on 16 Nvidia H100 GPUs plus four days on four GPUs, under $8,000, versus DeepMind DeepNash's estimated $3-4.5 million on 1,024 TPU nodes. Ataraxos learns only by self-play, using regularization to vary setups and a belief network to infer hidden pieces. The same method won Barrage series against top players, set Hanabi records, and beat PerfectDou and DouZero; at a 2025 worlds exhibition it won 38 of 40 games.
- Ataraxos posted an 85% effective win rate against champion Niemeijer.
- Training used 16 H100s for a week and cost under $8,000.
- DeepNash used 1,024 TPU nodes for months, estimated at $3-4.5 million.
- A belief network and regularization handle Stratego's hidden information.
- The method also led Hanabi and beat prior Dou dizhu systems.
From the original article
5 min read at the-decoder.com
The AI system Ataraxos has decisively beaten the best Stratego player of all time. The board game is a tough test for AI because both sides set up their pieces face down. Google Deepmind fell short in 2023 despite a multimillion-dollar budget.
By Maximilian Schreiner · the-decoder.com
A short excerpt; the article and its images belong to the publisher. Summary, key facts and assessment above are AI-written from the article and can be wrong.