MIT's New Game-Playing AI Beats Top Human Stratego Players

It is a system developed by MIT researchers that can defeat top-ranked human Stratego players. Stratego requires decisions under incomplete information, making it closer to real-world challenges than chess.

In Stratego you cannot see the opponent's pieces, so the AI must act under significant uncertainty. Chess and Go are games of complete information, which makes Stratego a tougher test of inference and reasoning.

No published information exists on commercial availability or licensing. However, this type of research is useful for regional AI teams building decision-support systems for high-uncertainty environments.

All involve incomplete information about the other side's intent and require balancing risk against reward. That is why this research may inform decision-support systems for negotiations and planning.