a4 Critical Position

1...Kh7

The rook endgame with a 5 versus 4 pawn configuration, involving passed a- and b-pawns (with attacking rook behind the passed pawn), is much more favorable to the superior side, than the analogous 4 versus 3 endgame. However, it is inaccurate to consider it generally winning; if the weaker side maintains a solid pawn structure, the draw is often the legitimate outcome.

Sam Shankland provides analysis of two such endgames in his recent methodically very valuable book "Theoretical Rook Endgames" (Quality Chess, 2023): Spassky-Torre (a-pawn) on pages 254-255 and Gelfand-Lagarde (b-pawn) on pages 256-257, 272.

In the Gelfand-Lagarde endgame with a b-pawn, despite numerous mistakes from both players, the game ended in Gelfand's favor. Shankland's analysis does not acknowledge these errors and concludes that "In general, a four on four kingside pawn structure should mean that pushing the pawn to the seventh rank will result in a victory " (p. 257). He extends this principle to a-pawns and formulates the following rule: “When the attacking side has their rook in front of an outside passed pawn, transitioning into a rook on the eighth and pawn on the seventh scenario tends to win if each player has four pawns on the other side of the board” (p. 272).

However, this rule even in its mild form (“tends to win”) is not correct also in case of a-pawn. In the case of the a-pawn - Spassky-Torre endgame, the victory by pushing the a-pawn to the seventh rank was not a legitimate result. Shankland considers the game from move 46 onwards where the position of the stronger side is indeed won. But until the 39th move (39...Rd2?) the weaker side missed several opportunities to draw. Shankland writes (p. 254) that he came across this endgame while making a video for Shereshevsky's book "Endgame Strategy" (New In Chess, 2022). Unfortunately, Shereshevsky's evaluation in this regard contains several errors (p. 119-123). Thus, the unjustly criticized move 36...h5 is indeed a good one and is very safe. Similarly, structures such as the f6-pawn, criticized heavily, can indeed be healthy and lead to a draw if handled correctly. This structure with f6-pawn arose in Carlsen-So 2021, where Carlsen could draw but he did not find the right path to it. The game ended with So's victory following numerous result-changing mistakes by both players, particularly notable during moves 5-7.

Attached are my detailed analyses of these endings. Further discussions will be initiated in separate threads.