Knight endgames are hard

Font Size
Bobby Ang

Chess Piece

Last Thursday we studied the Knight-and-Pawn Endgame in Alekseenko vs. Giri. Nowadays we have the benefit of the endgame tablebases to assist us in the analysis of these endgames. As BW readers know 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.

Before the development of the tablebases, chessplayers had to work out the do’s and don’ts of endgame play with their own brains. Do you know that the theory of that specific knight and pawn endgame in Alekseenko vs. Giri was first brought to light by GM Eugene Torre in the crucial game Torre vs. Portisch from the 1982 Toluca Interzonal?

Torre went to Toluca with every hope of qualifying for the Candidates’ Matches, but started very badly, scoring four draws and a loss (from a winning position against Igor Ivanov, formerly Soviet who was already representing Canada then) for 2.0/5 and a place near the bottom of the tournament table.

After the loss Eugene reckoned that he had no more chances to qualify and, with nothing to lose, started playing in a carefree manner. Feeling no pressure at all he threw caution to the winds and shocked the chess world with a five-game winning streak against several of the top GMs in the world, starting with Lev Polugaevsky (USSR), the Cuban Amador Rodriguez, then Yugoslavia’s champion Krunoslav Hulak, the American champion and sensation Yasser Seirawan, and finally the tough Lebanese-turned-French champion Bachar Kouatly.

Suddenly three rounds before the end he was leading the field with 7.5/10 followed by Lajos Portisch (7.0/10) and three pre-tournament favorites with 6.5/10: Yasser Seirawan (USA), Artur Jussupow (USSR), and former world champion Boris Spassky (USSR).

In the 11th round he met Hungarian living legend Lajos Portisch. With the tournament leaders all tightly bunched it was obvious that there was no question of an early draw — they were both in for the fight of their lives.

Torre, Eugenio (2535) — Portisch, Lajos (2625) [E15]
Toluca Interzonal (11), 1982

1.d4 Nf6 2.c4 e6 3.Nf3 b6 4.g3 Ba6 5.b3 Bb4+ 6.Bd2 Be7 7.Bg2 Bb7 8.Nc3 0–0 9.0–0 Na6

Portisch was among the leading theoreticians of his day. He introduced this move for the first time in this game, and it has since become the main line. Previously everybody automatically played 9…d5.

10.Ne5 d5 11.cxd5 exd5 12.Bg5 c5 13.Rc1 Ne4 14.Bxe7 Qxe7 15.e3 Rfc8 16.Ne2 Nb4!

Tempting white to move his pawn to a3, where it becomes a target for the black queen.

17.a3 Na6 18.Nf3 cxd4 19.Rxc8+ Rxc8 20.Nfxd4 Qxa3 21.Nf4 g6 22.Qa1! Qxa1 23.Rxa1

Intending to win back his pawn with Bf1.

23…Rc5! 24.Bf1 Ra5 25.Rxa5 bxa5 26.Bxa6 Bxa6 27.Nc6?

A little bit too flashy — he should have simply taken the d5–pawn. Now Portisch uncorks a small maneuver to get rid of his d5 weakness.

27…Nd2! 28.Nxa5 d4 29.exd4 Nf3+ 30.Kg2 Nxd4 31.f3 Kg7 32.Kf2 Kf6 33.Ke3 Ke5 34.Nd3+ Bxd3 35.f4+ Kd5 36.Kxd3 h5 37.b4 Nf3 38.h3 h4 39.gxh4 Nxh4 40.Ke3 f6

The time came for Eugene to seal a move. Elated with the fact that the draw is already well within reach he makes a mistake.


The move 41.Ke2 holds everything. 41…Nf5 (Black cannot advance with his king because then 41…Ke4 42.Kf2 Kxf4 43.Nc6 wins Black’s a-pawn) 42.Kd3 Nd4 43.Nc4! Ne6 44.Ne3+ Kd6 45.f5!= No kidding, these knight endgames are very difficult to play.

41…Nf5+ 42.Kd3 Nd4 43.b6 axb6 44.Nc4 b5 45.Ne3+ Kc5 46.f5! gxf5 47.h4 f4 48.Ng4 f5 49.Nh2 Nc6 50.h5

Now Eugene goes after the b-pawn in order to leave Black with doubled pawns on the same file.

50…Ne5+ 51.Ke2 b4 52.h6 b3 53.h7 Ng6?

The question mark comes from Portisch, who claims that 53…Nf7! wins. What is the difference between putting the knight on f7 and g6? In fact on g6 it performs the additional function of defending his f4–pawn. It turns out that on f7 a valuable check becomes available on g5. It appears that he is right, Black wins if Eugene defends the same way as he did in the actual game. However, there is still a way to a draw: 53…Nf7 54.Kd3 Kb4 55.Nf3 Ka3 56.Nd2 b2 57.Nc4+ Kb3 58.Nxb2 Kxb2 59.Ke2! we have a draw as Black’s king cannot get back in time to save his pawns.

54.Nf3 b2 55.Nd2 Kd4 56.Kf3 Kc3 57.Nb1+ Kd3 58.Kf2 Ke4

Now Black realizes the dilemma. If he tries to queen his b-pawn then 58…Kc2 59.Na3+ Kc1 60.Kf3 b1Q 61.Nxb1 Kxb1 62.h8Q Nxh8 63.Kxf4 draws, so he has to reconcile himself with the reality that the b-pawn must be given away for its counterpart white pawn on h7.

59.Nd2+ Kd5 60.Ke2

Game was adjourned for the second time here with White sealing the text move. Eugene was very distraught at the prospect of his qualifying chances melting away and came away from the board with a sinking feeling of hopelessness. GM Miguel Quinteros of Argentina, who was a close personal friend of Eugene, was present in Toluca as a correspondent for several big newspapers in his country, and did not hesitate to offer his services in the adjournment analysis.

They were both aware of the importance of the game and so really burned the midnight oil on this one. However, after several hours’ study they were forced to conclude that the endgame was lost. There was nothing left to be done but to catch a few hours sleep before the morning adjournment so that at least Eugene could meet the Hungarian legend with a clear head.

The game would not leave Eugene’s mind, though, and after sinking into an exhausted sleep the adjourned position broke into his dream with the knights and pawns dancing around the board. They were moving at a rapid pace all over the board and then, suddenly, a sequence of moves kept repeating itself in his dream and, subconsciously realizing its significance, he was shocked into wakefulness.

“Miguelito! Miguelito! It’s a draw! It’s a draw!” This call in the wee hours of the morning brought Quinteros running, and once again the feverish analysis started. Finally, with daybreak peaking from the waning darkness of the night they both sat back, smiling and happy. They had really found it.

60…Ke6 61.Kd3 Kf6 62.Kc2 Kg7 63.Nf3 Kxh7 64.Kxb2 Kh6 65.Kc2 Kh5 66.Kd2 Kg4 67.Nd4!

[67.Ke2? Kg3 68.Nd4 Kg2 69.Nf3 Ne7 70.Nd4 Nd5! 71.Nf3 (71.Nxf5? f3+) 71…Nc3+ and wins]

67…Ne5 68.Ke2 Nf7 <D>

Position after 68…Nf7

The position on the board more or less forcibly arises after the adjournment. They had concluded that it was lost because of 69.Kf2 Ng5 70.Kg2 (70.Ke2 Ne4 71.Kf1 f3 72.Ne6 Kg3 73.Ke1 f2+ 74.Ke2 f4) 70…f3+ Kf1 Kf4 and now the winning method is obvious — Black will bring his knight to g4 and king to e3 and force through the pawn. Here is how it will go 72.Kf2 Ne4+ 73.Kf1 Nf6 74.Kf2 Ng4+ 75.Kf1 Ke4 76.Nb5 Ke3 the sequence …Nh2+ followed by …f2 and queens cannot be prevented.

But this was the sequence Eugene saw — there WAS a way after all to prevent Black’s king from going to f4 and crossing over to the other side.


The dream sequence. Eugene’s knight is already on d4 and that is good — his King has to be on f2 when Black plays Ng5 otherwise he won’t be able to maintain his blockade on f3. If immediately 69.Kf2 Ng5 70.Kg2 (70.Ke2 Ne4 71.Kf1 f3 72.Ne6 Kg3 73.Ke1 f2+ 74.Ke2 Nd2) 70…f3+ 71.Kf1 (71.Kf2 Kf4 once the Black king gets to f4 it is curtains for White) 71…Kf4 72.Kf2 Ne4+ 73.Kf1 Nf6 74.Kf2 Ng4+ 75.Kf1 Ke4 76.Nb5 Ke3 Portisch wins.


[69…f3 70.Kf2 Kf4 71.Nxf5!=]


This is the same position as in 69.Kf2 Ng5, only now it is Black to play, and it is a draw.


If 70…f3 then 71.Ke3 preventing Black’s king from going to f4. Black is now in a bit of a zugzwang. If he moves his king then Nxf5. If he moves his knight then Nxf3. Take note that white’s king is two diagonal squares away from the black knight, meaning that it will take black a minimum of three moves before it can check the white king. Just a small tidbit to remember if you want to avoid any surprise checks.


The blockade on f3 holds. Portisch has to find a way to break it.

71…Nd2 72.Kf2 Ne4+ 73.Kg2 Nd2 74.Kf2 f3 75.Ke3!

The second part of the dream. Eugene’s king has to go to have access to e3 when the Black pawn advances to prevent …Kf4.

75…Nc4+ 76.Kf2 Nd2 77.Ke3 Nc4+ 78.Kf2

Portisch makes one last winning try.

78…Ne5 79.Ke3 f4+ 80.Kf2 Nd3+ 81.Kf1 Ne5 82.Kf2 Nd3+ 83.Kf1 Nc5 84.Kf2 Ne4+ 85.Kf1 Ng5 86.Kf2 Nh3+ 87.Kf1 Ng5 88.Kf2 Ne4+ 89.Kf1 Kg3 90.Nf5+ Kh3 91.Nd4 f2 92.Nf5! f3

[92…Ng3+ 93.Kxf2 Nxf5 94.Kf3 is a draw]

93.Nd4 Kg4 94.Nxf3! Kg3

[94…Kxf3 stalemate!]

95.Ke2 Nc3+ 96.Kf1 Nd5 97.Nd2 Nf6 98.Nf3 Ng4 99.Ke2 Ne5 100.Ne1! ½–½

Taking the knight is an immediate draw, the black king cannot move without losing the pawn, and if his knight moves then Nd3 would win the pawn.

With this very important draw GM Eugene tied with Lajos Portisch for 1st place in the 1982 Toluca Interzonal and qualified for the Candidates’ matches.


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.