This endgame is a classic, widely analyzed by numerous commentators. It was first examined by Kopayev in Shakhmatny Bulletin (No. 9, 1956, pp. 265–266), and later included in the first edition of Shakhmatnye Okonchaniya ("Chess Endings: Knight vs Bishop, Rook Endings", ed. Yu. Averbakh, Moscow, 1958, pp. 348–349). Interestingly, Averbakh omitted this analysis in the second Russian edition of the book, published in 1984.

In early 2012, I pointed out several errors in Kopayev’s analysis to Mark Dvoretsky. My corrections were incorporated into the second Russian edition of Dvoretsky's Endgame Manual (DEM, 2016, p. 248), as well as the fourth (2014) and fifth (2020, p. 203) English editions.

After nearly a decade away from chess analysis, I revisited this endgame in the summer of 2023. What prompted my return was Boris Gelfand’s excellent book Decision Making in Major Piece Endings (Quality Chess, 2020), in which he draws heavily on DEM and Korchnoi’s Practical Rook Endings (Ed. Olms, 2002, pp. 19–25).

This second examination of the game revealed further analytical oversights:

1.      Most significantly, my analysis demonstrates that Black's position is already lost after 41...Ra7. White wins with 42.e5!—a move previously overlooked by all analysts. After 42...Re7, White can defend the pawn from behind with 43.Re2!, a defense not possible after the inaccurate intermediate check 42.Rc8+?.

2.      In the line 42.Rc8+? Kh7 43.e5 Ra4? 44.Kg3 Re4 45.Re8 g5 46.hxg6+ Kxg6, the move 47.f4 is also winning. After 47...h5 (which DEM, p. 202, claims leads to a draw), White prevails with 48.g5!.

3.      After 42.Rc6? Ra3?, as played in the game, DEM (p. 203) correctly gives 43.e5! as a winning move. However, White can also win with 43.Rc8+, a line noted by Dvoretsky in the Russian edition (p. 248) but omitted in the latest English edition (p. 203).