Winnable positions that required even more moves were later discovered, however, and in 1992, FIDE abolished all such exceptions and reinstated the strict 50-move rule over the board.
[6][7] In correspondence chess under ICCF rules, the fifty-move rule only applies when more than seven pieces remain on the board; when seven pieces or fewer remain, a win or draw may be claimed with reference to endgame tablebases.
Such a position is sometimes termed a "cursed win" (where mate can be forced, but it runs afoul of the 50-move rule), or a "blessed loss" from the perspective of the other player.
[9] In retrograde analysis problems, castling also resets the fifty-move rule counter.
No claim needs to be made by either player, as the draw is mandatorily applied by the arbiter.
White is striving for the winning Philidor position while Black is employing the drawing Cochrane Defense and the "second-rank defense" (see rook and bishop versus rook endgame).
[16] Instead, the game continued for several more moves: The players agreed to a draw at this point because after 115.Kxf6 the position is a stalemate.
Instead, the game continued and Black resigned on his 142nd move after White had achieved a forced winning position even though the right to claim a draw was still in effect.
Another unusual event occurred in the Aeroflot Open of 2008 in a game[20] between Nguyễn Ngọc Trường Sơn and Maxime Vachier-Lagrave.
The opening day of the World Rapid Championship, 2024 witnessed a rare and bizarre end to a game because of 50 move rule.
German Grandmaster Alexander Donchenko made a great escape just by one move away from getting checkmated by Hikaru Nakamura of the USA, thanks to his use of the 50-move rule.
The arbiter came up with her notebook and verified that Donchenko was spot on, denying Nakamura a sure win and giving the German an unexpected draw.
The rules used at the 1883 London tournament reset the count if there was a capture or pawn move, but still started the count when the claim to apply the rule was made instead of going back to the last capture or pawn move.
[23] During the time periods when the fifty-move rule admitted exceptions, there were a number of revisions.
Harkness notes that "Some of these unusual positions have been established and accepted by FIDE", including two knights versus a pawn.
Ken Thompson's investigations in the 1980s using the Belle chess computer discovered numerous endgames winnable in more than 50 moves.
However, these often involved seemingly random moves that defied human comprehension or analysis, in situations that would hardly ever occur in real gameplay.
In 2008, the record was 517 moves (assuming optimal play by both sides) to make a piece capture or exchange that achieves a simpler and more obviously winnable sub-endgame, for a particular position involving a queen and knight versus a rook, bishop, and knight.