A Hacker News post treats chess as a concurrent system with interleaved turns, then extracts formal invariants. The author outlines a modeling approach that separates safety checks into state invariants—type correctness, one king per color, turn parity—and transition invariants that constrain how the board evolves each move. This framing mirrors techniques used in verification of distributed software and provides a template for future formalizations.

State invariants include TurnParity, which ties move number parity to the active color, and PreviousPlayerNotInCheck, restating the rule that a player must finish a turn out of check. Transition invariants such as MoveCountStrictlyIncreases, TurnAlternates, PieceCountNonIncreasing, SingleCapturePerMove, and ExactlyTwoSquaresChange encode the precise board changes expected in a basic move. These constraints also block illegal piece duplication and guarantee only one piece moves each turn, simplifying model checking.

The post notes that extensions like castling or en‑passant break ExactlyTwoSquaresChange, because four or three squares change respectively, while pawn promotion leaves piece count unchanged. Commenter James Corey adds a historical footnote about the 19th‑century clarification that a pawn cannot promote to a king. This precision helps AI engines obey every rule.