MARK RUBERY CHESS
The endgame that has had the greatest effect from the brute force approach of the computer is the ending K + 2B v K + N. In 1851 the endgame analysts Kling and Horwitz evaluated it as a draw on the grounds that the king and knight form a fortress that cannot be breached.
For more than a century this judgment remained uncontested and repeated in numerous endgame manuals down the years. What Kling and Horowitz had found was that the knight positioned on either b7 (and thus also b2, g2 and g7) was able to repel direct attacking attempts as it controlled the critical a5-d8 diagonal, and the noted endgame expert Cheron supported their analysis. As the power of the computer was beginning to make swift strides it was proposed to Ken Thompson of Bell Laboratories that K + 2B v K + N would be the ideal ending for the construction of a database. Thus on completion in 1983 that new database overturned 132 years of endgame theory that the ending, a few minor positions aside when Black can force an immediate draw, is always winning for the bishops. In his excellent book ‘Secrets of Minor-piece Endings’ John Nunn painstakingly goes through the computer analysis, showing that through a sequence of optimal moves the blockade can be dismantled and the knight and king (and this is the difficult part) are unable to reach another corner of the board to erect the Kling and Horowitz position.
A few years ago South Africa’s lone grandmaster, Kenny Solomon, reached this ending with the knight when playing the Italian GM, Luca Moroni. Kenny was able to escape with the draw via the ’50 move rule’ when the winning method was beyond the grasp of his opponent.
We pick up the game after 93 moves where the table bases declare it mate in 56 moves after White’s choice whereas with 94 Kf4! it was a mate in 30 (or more practically a win of the knight in half that number).
Moroni,l (2549) - Solomon,k (2419) [E68]
4th ad Gredine Open 2018 Ortisei ITA (3.4) 2018 94.Bd1 (94.Kf4! Ne1 95.Be3+ Kf1 96.Bb6 Nd3+ 97.Ke3 Nf2 98.Bf3 Nh3 99.Bd8 Ng1 100.Be4 Nh3 101.Bh4 Ng1 102.Bb7 Nh3 103.Kf3 Ng1+ 104.Kg3 Ne2+ 105.Kh2 Nf4 106.Ba6+ Ne2 107. Kh3 and the knight drops) 94...Ne1 95.Bh5 Ng2 96.Bb4 Kg3 97.Bc5 Kh4 98.Bf7 Kg3 99.Be6 Ne1 100.Bd7 Ng2 101.Bd6+ Kf2 102. Bc6 Ne1 103.Be8 Ng2 104.Bh5 Ne1 105.Bc5+ Kg3 106.Kd4 Kh4 107.Bf7 Kg3 108.Ke3 Ng2+ 109.Kd2 Nf4 110.Bd6 Kf3 111.Be8 Ke4 112. Bc6+ Nd5 113.Bc5 Ke5 114.Kd3 Nf4+ 115.Kc4 0.5-0.5 as the last pawn left the board 50 moves ago
‘Although Kling and Horowitz did not evaluate the diagram correctly, they nevertheless found the defender’s basis for prolonged resistance, and that is an achievement in itself.’ –John Nunn