Endgame Theory

Many significant chess treatises, beginning with the earliest works, have included some analysis of the endgame. Lucena's book (c. 1497) concluded with 150 examples of endgames and chess problems.[85]

The second edition (1777) of Philidor's Analyse du jeu des Échecs devoted 75 pages of analysis to various endgames.[86] These included a number of theoretically important endings, such as rook and bishop versus rook, queen versus rook, queen versus rook and pawn, and rook and pawn versus rook. Certain positions in the endings of rook and bishop versus rook, rook and pawn versus rook, and queen versus rook have become known as Philidor's position. Philidor concluded his book with two pages of (in the English translation), "Observations on the ends of parties", in which he set forth certain general principles about endings, such as: "Two knights alone cannot mate" (see two knights endgame), the ending with a bishop and rook pawn whose queening square is on the opposite color from the bishop is drawn (see wrong rook pawn#Bishop and pawn), and a queen beats a bishop and knight (see pawnless chess endgame#Queen versus two minor pieces).[87]Staunton's The Chess-Player's Handbook (1847) includes almost 100 pages of analysis of endgames.[88] Some of Staunton's analysis, such as his analysis of the very rare rook versus three minor pieces endgame, is surprisingly sophisticated. At page 439, he wrote, "Three minor Pieces are much stronger than a Rook, and in cases where two of them are Bishops will usually win without much difficulty, because the player of the Rook is certain to be compelled to lose him for one of his adversary's Pieces. If, however, there are two Knights and one Bishop opposed to a Rook, the latter may generally be exchanged for the Bishop, and as two Knights are insufficient of themselves to force checkmate, the game will be drawn." Modern-day endgame tablebases confirm Staunton's assessments of both endings. Karsten Müller and Frank Lamprecht, Fundamental Chess Endings, Gambit Publications, 2001, p. 403. ISBN 1-901983-53-6. Yet Reuben Fine, 94 years after Staunton, erroneously wrote on page 521 of Basic Chess Endings that both types of rook versus three minor piece endings "are theoretically drawn." Grandmaster Pal Benko, an authority on the endgame and like Fine a world-class player at his peak, perpetuated Fine's error in his 2003 revision of Basic Chess Endings. Reuben Fine and Pal Benko, Basic Chess Endings (Algebraic Edition), McKay Chess Library, 2003, p. 524, ISBN 0-8129-3493-8. Grandmaster Andrew Soltis in a 2004 book expressly disagreed with Staunton, claiming that the rook versus two bishops and knight ending is drawn with correct play. Andrew Soltis, rethinking the chess pieces, Batsford 2004, p. 84. ISBN 0-7134-8904-9. At the time Benko and Soltis offered their assessments (in 2003 and 2004, respectively), endgame tablebases had already proven that Staunton was correct, and that Fine, Benko, and Soltis were wrong, although the ending can take up to 68 moves to win. Müller and Lamprecht, p. 403.

Staunton's conclusions on these endgames were, however, anticipated by the British master George Walker, who wrote in 1846 (and perhaps earlier):Although the two Bishops and Kt win, as a general proposition, against Rook, yet the two Knights with a Bishop cannot expect the same success; and the legitimate result of such conflict would be a draw. The Bishops, united, are stronger than the Knights, as they strike from a greater distance. When the two Knights are left with a Bishop, the Rook has also the chance of exchanging for the latter, which can hardly be avoided by his adversary, and the two Knights, alone, have not the mating power.[89]

In 1941 Reuben Fine published his monumental 573-page treatise Basic Chess Endings, the first attempt at a comprehensive treatise on the endgame.[90] A new edition, revised by Pal Benko, was published in 2003.[91]Soviet writers published an important series of books on specific endings: Rook Endings by Grigory Levenfish and Vasily Smyslov,[92] Pawn Endings by Yuri Averbakh and I. Maizelis,[93] Queen and Pawn Endings by Averbakh,[94] Bishop Endings by Averbakh,[95] Knight Endings by Averbakh and Vitaly Chekhover,[96] Bishop v. Knight Endings by Yuri Averbakh,[97] Rook v. Minor Piece Endings by Averbakh,[98] and Queen v. Rook/Minor Piece Endings by Averbakh, Chekhover, and V. Henkin.[99] These books by Averbakh and others were collected into the five-volume Comprehensive Chess Endings in English.In recent years, computer-generated endgame tablebases have revolutionized endgame theory, conclusively showing best play in many complicated endgames that had vexed human analysts for over a century, such as queen and pawn versus queen. They have also overturned human theoreticians' verdicts on a number of endgames; for example by proving that the two bishops versus knight ending, which had been thought drawn for over a century, can be a win for the bishops (see pawnless chess endgame#Minor pieces only and Chess endgame#Effect of tablebases on endgame theory).

1. Several important works on the endgame have been published in recent years, among them Dvoretsky's Endgame Manual,[100] Fundamental Chess Endings by Karsten Müller and Frank Lamprecht,[101] Basic Endgames: 888 Theoretical Positions by Yuri Balashov and Eduard Prandstetter,[102] Chess Endgame Lessons by Benko,[103] and Secrets of Rook Endings[104] and Secrets of Pawnless Endings by John Nunn.[105] Some of these have been aided by analysis from endgame tablebases.