We continue our discussion from where we left off last Tuesday — nowadays we have the benefit of the endgame tablebases to assist us in the analysis of endgames. As explained then an endgame tablebase is a computerized database that contains precalculated exhaustive analysis of chess endgame positions. They are generated by working backwards from a checkmated position. Thus, the tablebase acts as an oracle, always providing the optimal moves.
By 2005 all chess positions with up to six pieces (you include the Kings when you count pieces, therefore KRP vs KRP is considered a six-piece ending) had been solved. By August 2012, tablebases had solved chess for every position with up to seven pieces.
The solutions have profoundly advanced the chess community’s understanding of endgame theory. Some positions which humans had analyzed as draws were proven to be winnable and some positions thought to be winnable were proven to be drawn.
Before the development of the tablebases, chessplayers had to work out the do’s and don’ts of endgame play with their own brains. Last Tuesday I showed you the Torre vs. Portisch game which illustrated a critical idea in a difficult knight+pawns vs. knight endgame. Here is another Torre contribution to the theory of endgames.
The most important chess events of the year bring out the very best in the players. Eugene Torre won the 1972 Asian Zonal and represented the region in the 1973 Interzonals from which the top 6 players will go on to the Candidates’ Matches. The final winner of the Candidates’ matches will challenge Bobby Fischer for the world title. You will recall that ultimately Bobby Fischer refused to defend his title and the winner of the Candidates,’ Anatoly Karpov, was declared world champion in 1975.
Anyway the 36 interzonal qualifiers were split into two. Half of them played in Petropolis, from which Henrique Mecking (Brazil), Lev Polugaevsky (USSR) and Lajos Portisch (Hungary) advanced to the Candidates, and the other half in Leningrad, where Viktor Korchnoi (USSR), Anatoly Karpov (USSR) and Robert Byrne (USA) dominated.
Eugene played in Leningrad. One of the favorites, Bent Larsen, started off the tournament with 3/3, defeating Josip Rukavina (Rumania), Ivan Radulov (Bulgaria) and Jan Smejkal (Czechoslovakia) in impressive fashion. Everyone expected him to dust off Eugene Torre, not yet even a grandmaster at that time and a debutante in this level of competition, but Torre held tight in an endgame which was widely thought to be an easy win.
Larsen, Bent (2620) — Torre, Eugenio (2430) [A27]
Leningrad Interzonal (4), 1973
1.c4 g6 2.Nf3 Bg7 3.Nc3 e5 4.d4 exd4 5.Nxd4 Nc6 6.Nxc6 bxc6 7.g3 Ne7 8.Bg2 0–0 9.0–0 d6 10.Bg5 f6 11.Bd2 Be6 12.Qa4 Bd7 13.c5!
Torre: A typical Larsen move. The idea is to break Black’s pawns in the center, at the same time eyeing to bring his queen to the kingside.
13…d5 14.e4 f5 15.exd5 cxd5 16.Qh4
16…Bf6 17.Bg5 Bxg5 18.Qxg5 c6
Torre: White has a slight edge in space and has strong pressure along the open king’s file and also the half open queen’s file. White’s strategy is to place both his rooks along two open files and then try the thematic advance b2–b4–b5 splitting Black’s center pawns.
19.Rfe1 Rf7 20.Re5
During the game GM Eugene suddenly saw the possibility of 20.Nb5 but then determined that 20…cxb5 21.Rxe7 Rxe7 (21…Be6! is even stronger) 22.Bxd5+ Kg7 23.Bxa8? is refuted by 23…Re1+, winning White’s queen.
20…Kg7 21.Rd1 f4!
Torre: Black’s only chance to get some counterplay, otherwise White plays Qf4 followed by b2–b4–b5 with crushing initiative.
22.gxf4 h6 23.Qh4 Qf8
With the idea of Nf5.
[24…Rxf4? 25.Rxe7+; Better is 24…Nf5 25.Qd3 Rb8]
[26…Nf5 is better]
[27…Nf5! 28.Ra4 Nh4 29.Rxa5 g5 Black has an initiative going on]
28.Ra4 Qb8 29.b3 Qa7 30.b4 Qb8 31.a3 axb4 32.axb4 Nf5 33.Bh3 Rb7 34.Bxf5 Bxf5 35.Qa3 Ra7 36.Nd4 Rxa4 37.Qxa4 Ra7 38.Qb3 Be4
I hope you noticed that Black is threatening …Ra1 mate.
39.f3 Ra1+ 40.Kg2 Rb1 41.Qa4 Rb2+ 42.Kg3 g5!
Torre: Again Black found the only move to complicate the game. To retreat the bishop loses at once to Re7+ and Qxc6. The game was adjourned at this stage with Larsen sealing the move.
Torre: The best move to retain winning chances. 43.fxe4? would lead to a draw after 43…gxf4+ 44.Kxf4 Rf2+ 45.Nf3 Rxf3+ 46.Kxf3 Qxe5.
43…gxf4+ 44.Kxf4 Qf8+ 45.Nf5
[45.Ke3? Rb3+ 46.Nxb3 Qxf3+ 47.Kd2 Qd3+ with at least a draw]
45…Bxf5 46.Rxf5 Qe7
[46…Rxb4+? 47.Ke5 Qe7+ (47…Qg7+ 48.Ke6 Qg6+ 49.Rf6 Qg8+ 50.Rf7+ Kh8 51.Ke7 the game is over) 48.Qe6 Qg7+ 49.Kxd5 Rb7 (49…Qd4+ 50.Kc6 Black has no more checks) 50.Rf6 and the threat of Qf5+ is decisive]
47.Rf6! Rxb4+ 48.Kf5 Rh4 49.Qe6 Rh5+ 50.Kf4 Qc7+ 51.Qd6 Qg7 52.h4!
Torre: This is White’s winning move as we had seen in our analysis. Larsen played this move very fast and we thought he had analyzed this variation completely. He walked around the stage and smiled at some of his friends.
52…Rxh4+ 53.Kf5 Rh1
Torre: We saw in our analysis that this move would lose. But there is nothing else that could give White some problems. The idea is to place the rook on e1.
[54.c6! Re1 (54…Qg5+ 55.Ke6 Re1+ 56.Kd7 Qg7+ 57.Kd8 Ra1 58.c7 is decisive) 55.Rg6! Qf7+ (55…Re5+ 56.Kf4 White wins) 56.Qf6 Qxf6+ 57.Rxf6 Rc1 58.Ke5 Kg7 59.Re6 White wins because his opponent’s king is cut off from the c6–pawn]
54…Qxf8 55.Rxf8 Kg7?
[55…Rc1! holds the draw. For example 56.Rc8 h5 57.c6 d4 58.Ke4 h4 59.Kxd4 h3 60.Rc7+ Kg6 61.Rc8 h2 62.Rh8 h1Q 63.Rxh1 Rxh1 64.Kd5 draw]
[56…h5 57.Rxd5 h4 58.Rd7+ Kh6 59.c6 Rc1 60.Rd4 h3 61.Rh4+ Kg7 62.Rxh3 Rxc6 This endgame is a book draw]
57.Rd7+ Kf8 58.Kf6 Ke8 59.Rxd4 Rc1 60.Rd5 Rf1 61.Rf5
Larsen is now completely winning but accuracy is still required.
61…Kd7 62.f4 h5 63.Kg5 h4 64.Kxh4 Ke6 65.Re5+ Kf6 66.Kg4 Rc1 67.Rh5 Kg6 68.Rd5 Kf6 69.Kf3 Rc3+ 70.Ke4 Rc4+ 71.Ke3 Ke6 72.Rh5 Kd7 73.Kf3 Ke7 74.Kg4 Rc1 75.f5 Kf8
76.Rh8+ Ke7 77.Rh7+ Ke8 78.Rc7
The way to win this endgame is to transpose from a R+2P versus R endgame to an easily winning R+P versus R. He could accomplish this by 78.Kg5! Rxc5 79.Kg6 with the idea Rh8+. White has a simple win because the black king is on the long side. 79…Rc1 80.Rh8+ Kd7 81.f6 Rg1+ 82.Kf7 Rf1 83.Rh6 Rg1 84.Rg6 Rf1 85.Kf8 Ra1 86.f7 Rf1 87.Rg7 (87.Kg7?? Ke7=) 87…Ra1 88.Kg8 the end. The move chosen by Larsen does not give anything away yet. That comes later.
78…Kd8 79.Rc6 Kd7 80.Rd6+ Ke7 <D>
Torre: SuperGM Larsen again played very fast at this point and did not realize that his last move was a serious mistake. Correct is 81.Re6+! Kd7 (81…Kf7 82.c6 Rf1 83.Kg5 Rf2 84.Rd6 Black would have had no alternative but to resign.) 82.c6+ Kc7 83.Kg5 Rg1+ 84.Kh6 Rg2 85.f6 Rf2 86.Kg7 Rg2+ 87.Kf8 Kd8 88.Re7 Rh2 89.c7+ Kc8 90.f7 Rg2 91.Ke8.
81…Kf7 82.c6 Kg6 83.Kf3 Re1
This is White’s problem. If he advances the c-pawn then 84.c7 Rc1 forces him to give up one of the pawns with a dead draw. The Black rook cuts off his king from the queenside and also on the kingside Larsen’s king cannot approach his own pawn. Now it is clear that White cannot strengthen his position.
84.Kf4 Re2 85.Rd5 Rc2
[85…Kxf6? is erroneous: 86.Rc5 Re8 87.Rf5+! Kg6 (87…Ke7 88.Re5+ Kf7 89.Rxe8 Kxe8 90.Ke5+–) 88.Re5! Rc8 89.Re6+ Kf7 90.Ke5 wins]
86.Rd6 Re2 87.f7+ Kxf7 88.Kf5 Ke7 89.Rd7+ Ke8 90.Kf6 Re1 91.Rd5 Rc1 92.Rd6 Rf1+ 93.Ke6 Re1+ 94.Kd5 Rd1+ 95.Kc5 Rxd6 96.Kxd6 Kd8 ½–½
So, finally, the “Great Dane” was halted. Torre got a tremendous ovation from the gallery and received flowers from his new fans.
GM Joey Antonio told me a story of the 1998 Elista Olympiad, held in Kalmykia, Russia (take note that this is 25 years after the Leningrad Interzonal). During the rest day the Philippine team went to a nearby marketplace and, to their surprise, the market vendors recognized Torre, and remembered him as the one who stopped Larsen. Yes, Russians take their chess seriously.
Bobby Ang is a founding member of the National Chess Federation of the Philippines and its first Executive Director. A Certified Public Accountant, he taught accounting in the University of Santo Tomas for 25 years and is currently Chief Audit Executive of the Equicom Group of Companies.