Combinatorial Games in Lean

A new project called Combinatorial Games in Lean sounds, at first glance, like peak academic energy: researchers are teaching a proof-checking computer system to understand two-player games where everyone sees everything, nobody gets lucky, and someone always loses because they run out of moves. Think Nim and tic-tac-toe, not poker or regular chess. On paper, it’s about turning deep game ideas into something a machine can verify step by step. In the comments, though? The real sport was nitpicking the definition and joking about the edge cases.

The strongest vibe was a mix of genuine admiration and classic internet side-eye. One commenter gave the wholesome suggestion to add a wishlist or project board, basically saying: nice brainy project, now make it easier for the rest of us to stalk the roadmap. But then the thread swerved into deliciously nerdy chaos when another commenter pounced on the line saying chess is a non-example because it can end in a tie. Their reaction was basically: and yet somehow… which is exactly the kind of half-finished roast that lands harder because everyone knows where it’s going. It’s a tiny comment, but it carries big "chess players are never beating the draw allegations" energy.

There was also a stray but spicy detour into so-called "quantum games," with one commenter recalling a paper claiming that, under certain assumptions, making games quantum doesn’t magically give players an advantage. Translation: even the futuristic version might still be less exciting than the comment section. So yes, the project is serious math. But the crowd turned it into a mini-drama about definitions, missing features, and whether chess just got subtweeted by a theorem.