Shakmaty Bereolos - The Official Chess Site of Peter Bereolos

3/29/20 - von Brühl-Philidor, London 1783

One thing I've enjoyed as part of my BCE project is learning more about the history of chess. However, BCE was published in 1941 so there are nearly 80 years after that to explore. As my recent analysis of the Kasparov-Karpov ending showed (This time GM Timman fully concurred with my analysis. For a player who came of age at the time that Kasparov, Karpov, and Timman were the kings of chess, it is thrilling to be able to make a contribution to a position associated with those 3 names), there are still more secrets to be revealed in more recent. Since it looks like sources of new OTB material may be drying up for awhile, I'm going to periodically pick out other historical endgames to analyze.

Before moving forward in time, I'm going to first go into the deep past. The name Francois Philidor is heavily linked with rook and pawn endings and one of the first rook endings in my database is one by Philidor. This game took place as part of a 3-board blindfold simul in which Philidor scored 2 wins and a draw. His opponent, Hans Moritz von Brühl, was described as one of the best players in London in a newspaper account that I found and his Wikipedia page describes him as one of the best players of his time. After 40...Rxf4

41.Rxd5 Rf3 42.Rd8 It was probably better to stop the connected passers from getting started by putting the rook on the e-file. 42.Re5 Kf6 43.h5 Rd3 44.h6 Rxd4 45.Re8 Rd7 46.Kg3 Rh7 47.Kf4 Rxh6 48.Ra8 and White is holding 42...Rd3 43.d5 f4 44.d6 Consistent, but 44.Re8 is a sensible alternative 44...Rd2+ 45.Kf1 Kf7 46.h5 The drawing window is already starting to close, but it is still there with 46.d7 e3 47.Re8 Rxd7 48.Re4! Rd1+ 49.Ke2! Rd2+ 50.Ke1! Rf2 51.h5! Kg7 52.Re6 Kh7 53.h6 Rh2 54.Rf6 46...e3 47.h6? The final mistake. White absolutely had to prevent ...f3 47.Rd7+! Ke6 48.Re7+ Kxd6 49.Re4 Rf2+ 50.Kg1! (50.Ke1? Kd5! 51.Re8 Rh2 and Black picks up both h-pawns) 50...Kd5 51.Re8 and White is holding. Black even loses with 51...Kd4? 52.h6! because White threatens to queen with check.47...f3 0-1

3/28/20 - Bereolos-Goldin, 2002 Knigs Island Open

I've posted my draw with Alexander Goldin at the 2002 Kings Island Open to the GM games section. The tournament situatation heavily influenced my play in this game. I deliberately went for a slightly inferior position out of the opening because I was confident I could hold a draw and secure a very large prize.

3/27/20 - Shen-Gauri, 2019 CCSA Fall IM

The winner of the IM norm section of the CCSA's Fall invitational was Christopher Shen, who made his first IM norm. A couple of endings played an important role in his result. In the third round, he squeezed out a win from a knight ending against Gauri Shankar after 49...Nxf2

White has a number of pluses here, in particular the outside passed pawn. His king can also get to the center faster and Black has some kingside weakness. Still, Black ought to be able to hold because of the limited material. 49...Nxf2 50.Kg2 Ne4 51.b4 Nd6 52.Kf3 Kf8 53.Kf4 Ke8 54.Ke5 Kd7 55.Ne3 Kc6 56.Kf6 Kb5 57.Ke5 Ne8 58.Nd5

58...Kc4? It looks like Black has constructed a fortress, but White shows that it can be broken. This was the last chance to draw by activating the knight 58...Ng7 59.Kf6 (59.g4 hxg4 60.hxg4 Ne8 61.g5 Kc4 62.Nf6 Ng7 63.Ne4 Kxb4 64.Kf6 Ne6) and the last White pawn goes 59...Nf5 60.Kxf7 Nxg3 61.Kxg6 h4 62.Kg5 Ne4+ 63.Kxh4 Nf2 64.Kg3 Nxh3! 65.Kxh3 Kc4!= 59.h4! Kb5 Now it is too late for 59...Ng7 60.Ne3+ Kxb4 61.Kf6! and the White king mops up the pawns 60.Nf6! Ng7 61.Ne4! f5 Trying to stop the king invasion leads to a lost pawn ending 61...Ne8 62.Nd6+ Nxd6 63.Kxd6 Kxb4 64.Ke7 f5 65.Kf6+- 62.Nd6+ Kxb4 63.Kf6 Kc5 64.Kxg7 Kxd6 65.Kxg6 Ke5 66.Kxh5! Kf6 67.Kh6! f4 68.gxf4! Kf5 69.h5 Kf6 70.f5 Kf7 71.Kg5 1-0

3/25/20 - BCE-180a, Santasiere-Kashdan, Boston 1938

This week's BCE position is somewhat of a soft correction. In the line in question, I think Fine was looking to point out the wrong idea (capturing the Black f-pawn with his King on g5 instead of e5). However, the move that starts the line (3.Kg5) does not spoil the win, it is only the bishop check on the next move that is the mistake. Euwe and Hopper give the bishop check a question mark in their book, so I think it is fair game as a correction.

The game between Anthony Santasiere and Isaac Kashdan at the 1938 American Chess Federation Congress was a marathon struggle that was adjourned 5 times! Santasiere finally forced a winning bishop ending after 104.Bxd3+

Despite the limited material, the outside passed pawn is decisive. 104...Kb7 105.Bf5 Be8 106.Be6 Bg6 107.Ke3 Kc7 108.Kd4 Kd6 109.Bc4 Bf5 110.Bd3 Bh3 110...Bc8 111.a6 Ke6 112.b5 cxb5 113.a7 Bb7 114.Be4 111.a6 Kc7 112.Kc5 Bg2 113.Be4 White forces through connected passers with 113.b5 Bxf3 (113...cxb5 114.Be4 Kb8 115.Kb6!) 114.b6+! Kb8 115.Bf5 Be2 116.a7+! Kb7 117.Be4 Bb5 118.Kd6 f5 119.a8Q+ Kxa8 120.Kc7! 113...Bf1 114.Bxc6 White could liquidate to a winning pawn ending with 114.b5 Bxb5 115.a7! Kb7 116.Bxc6+ Bxc6 117.a8Q+ Kxa8 118.Kxc6! 114...Bxa6 115.b5! Surprisingly, White can't wait with this move. Instead, 115.Be4? Be2 116.b5 f5! 117.Bd5 Bd1 118.b6+ Kb8 119.Kd6 Bc2 120.f4 Bb1 121.Ke5 Bc2 122.Be6 Kb7 123.Bxf5 Bd1 124.Be6 Kxb6 (or 124...Bc2 where unlike the game 125.Kd4 would not hit the bishop) 125.f5 Kc7 126.f6 Bh5 and Black is in time 115...Bc8 116.b6+ Kb8 117.Kd6 Bh3 Black doesn't have time to transfer his bishop to the longer diagonal as 117...Bf5 118.Ke7 immediately loses the f-pawn 118.f4 Bg4 This is the starting position of BCE-180a except Fine has the White bishop on d5 instead of c6. Irving Chernev analyzed the ending in the October 1938 edition of The Chess Review and also starts with the BCE position. He states 120 moves had been played, so maybe a couple of moves got left out when the game was entered into databases. However, in Reinfield's report in the August 1928 edition of The Chess Review, he says the game went 127 moves, which agrees with the database. I was unable to find a print version of the game score, so it is another mystery. 119.Ke7 f5 120.Kf6 Bh3 120...Kc8 If Black tries to keep the bishop on g4 in order to have ...Bd1 after White plays Be6xf5, White wins with 121.Bg2 with the idea of Kg5 and Be4 121.Ke5 121.Kg5 Kc8 is the line in the BCE correction, but with the bishop on c6 White doesn't even have the move Be6+? 121...Kc8 122.Bf3 Kb8 123.Bd5 Kc8 124.Be6+ Kb7 125.Bxf5! Bf1 126.Be6 Bd3 Black also falls short if he takes the pawn as demonstrated in the following lines by Chernev 126...Kxb6 127.f5! Bd3 (127...Kc7 128.f6! Kd8 129.Bf7! Bd3 130.Kf4! Kd7 131.Kg5! Kd6 132.Kh6 Ke5 133.Kg7! Bb5 134.Bb3 Be8 135.Bc2) 128.f6! Bg6 129.Kd6 Bh5 130.Ke7 Kc5 131.Bf7 Be2 132.Bg6 Bc4 133.Bf5 Kd4 134.Be6!+- 127.Kd4! White will remain two pawns up after the bishop moves by 128. Kc5! so Black resigned.1-0 Santasiere won a special prize for hardest fought game for this effort.

3/23/20 - Ofek-Bereolos, 2012 Kings Island Open

In the opening round of the 2012 Kings Island Open, I dropped a pawn in the middle game, but put up a good defense and was even pressing for the win still a pawn down in the endgame after 50...Kh3

51.c4? The not so obvious problem with this move is that Black gets the d3 square for his bishop. Better was 51.Ne4 and Black doesn't have a good way to improve further 51...bxc3! 52.bxc3 Bd3! 53.c4 Bxf5? The bishop is dominating the Ng3, so Black should dislodge Nf5 53...Nh4-+ with an easy win 54.Nxf5! g3! 55.Ne3? 55.Nxg3! draws Kxg3! 56.f5 Kf4 57.f6 Ne5 58.Kc2 Kf5 59.b4 a4 60.f7 Nxf7 61.Kb2 Ne5 62.Ka3! 55...Nd4! 56.c5 g2? very bad. Black wins with the simple deflection 56...Nf5!-+ 57.Nxg2 Kxg2 58.f5 Nxf5 59.Kd2 Kf3 60.Kd3 Kf4 61.Kc4 Ke5 62.Kb5 1/2-1/2

3/22/20 - Zhou-Jacobson, 2019 CCCSA Fall IM

The knight ending of the game between Liran Zhou and Aaron Jacobson at the Charlotte Chess Center's Fall IM norm event had a surprising outcome. After 46.Nxd2 (note, I am using the move numbers from issue 1309 of the Week in Chess, but there was obviously some problem with the move transmission on moves 14, 15, and 16)

It doesn't seem that there should be much left to play for as the material is even and the pawns are balanced. However, the White king is a bit hemmed in, so Black plays on. 46...Kf6 47.Ne4+?! a step in the wrong direction. Better is 47.Nf1 intending g3 47...Nd4 48.Kg1 47...Ke5 48.Nd2 f5 49.Nf1 Kf4

50.Nd2? Black has made enormous progress, but White seems to be holding a fortress with 50.Kh1! Ng3+ 51.Kg1! (51.Nxg3? hxg3! takes the key squares f2 and h2 under control 52.Kg1 Ke3 53.Kf1 Kd2 54.Kg1 Ke2 55.Kh1 Kf2 56.h4 Kf1 57.f4 Ke2 wins both the f and h pawns) 51...Nxf1 52.Kxf1 Ke3 53.Ke1 f4 54.Kf1! Kd2 55.Kf2 and White can't be driven from f1-f2-e1 50...Ke3! 51.Nc4+ Kf2! 52.Nb2 Nf4! 53.Nd1+ Ke2 54.Nc3+ Kf1 55.g4 55.g3 Ne2 -+ 55...Kf2! 56.Nb5

56...Ne2? 56...Nd5! 57.gxf5 gxf5! 58.f4 Kf3! 59.Nd6 Ne3! taking away g2 from the White king 60.Kg1 Kg3 57.Nd6? An inaccurate move order. White reaches the drawn position that occurs in the game with 57.gxf5! gxf5 58.f4! Kf3 59.Nd6! Nd4 60.Nc4 Kxf4 61.Kg2! 57...Nd4? It is hard to figure why Black didn't grab more space with 57...f4 as 58.Nf7 Nd4 should be an easy win 58.gxf5 gxf5 59.f4! Kf3 60.Nc4 Kxf4 61.Kg2! Ne6 62.Nd6 Ng5 63.Ne8? This puts the knight out of play. White should hold the drawing by bringing the knight back around with 63.Nc4= 63...Ke3? Black should keep the knight on a bad track with 63...Ke5 64.Nd6! Kf4 64...f4 65.Nf5+ so Black has to go back

65.Ne8? Repeating the position, but this time Black finds the win. 65.Nc4= 65...Ke5 66.Ng7 f4 67.Ne8 f3+ 68.Kh2 Kf4 69.Nf6 Ke3 70.Nd5+ Kd2 71.Nf4 Ne6 72.Nd5 f2 73.Kg2 Ke2 74.Nf4+ Nxf4+ 75.Kh2 Kf3 avoiding the stalemate trap 75...f1Q? 0-1

3/21/19 - Bereolos-Brown, 2019 US Masters

I came very close to a repeat performance of an opening round upset at last year's US Masters. Against one of the newest US grandmasters, Michael Brown, I was slightly worse, but he made a number of mistakes during the time scramble and I emerged with a technically winning position. However, he put up a tough defense and I finally lost patience during the increment phase and allowed a draw.

3/18/20 - BCE-401b, Saemisch-Treybal, Teplitz-Schönau 1922

This week's BCE position is rook ending between Fritz Saemisch and Karl Treybal from the 1922 tournament at Teplitz-Schönau. Both players landed in the bottom half of the standings well behind the co-winners Reti and Spielmann. After 62.Ra7

Treybal went for the rook ending with 62...Nxd5 rather than putting his knight in a very passive position with 62...Nc8 63.Rd7 hxg4 63.cxd5 hxg4! The starting position of BCE-401b. Like Fine, the tournament book considered the position winning for White. 64.Ke5 Rh6! The immediate 64...g3? costs a vital tempo as after 65.Ra3 White threatens to take the pawn with check. 65.Ra4 Kf8? 65...g3! is the subject of the BCE correction. The only alternative considered by the tournament book is 65...Rh1? 66.f6+ Kg8 67.Rxg4+ Kf8 68.Ra4 Re1+ 69.Kd6 and White wins 66.d6 Ke8 67.Ra8+! Kd7 68.Ra7+! Kd8 68...Ke8 69.Re7+ Kf8 70.f6 69.Rxf7 Rh1 70.Rg7 Rg1 71.f6 Re1+ 72.Kd5 Rd1+ 73.Kc6 Ke8 74.d7+ Kf8 75.Re7 b5 76.Re8+ Kf7 77.d8Q Rxd8 78.Rxd8! Kxf6 79.Kd5 1-0

3/11/20 - BCE-605a

This week's BCE position deals with the very esoteric endgame of queen versus two knights. Examining the history of the position revealed some curiosities. Fine only cites Handbuch with no year given. This is a reference to the famous Handbuch des Schachspiels, but that compendium of chess knowledge had 8 editions spanning from 1843 to 1916. I checked the first and last editions and they have the following position

Unlike BCE, the queen is on c4 instead of e2. Interestingly, the repositioning of the queen changes the assessment of the position. Handbuch position is winning for White whereas the BCE position is a draw. Both editions of the Handbuch give the same main line. 1.Qe6? 1.Qc7! cutting off the king is the only winning move. Strangely, in the revised edition Benko (with the queen on e2) changed Fine's starting move to 1.Qe2-a6. After the text, we are back to the position analyzed in both the Handbuch and by Fine. 1...Kg7! 2.Kf3 Nh7? In the first edition of the Handbuch this mistake is given as seems best By the time the final edition came out, the note was change to give the following alternative, which leads to a draw 2...Nh8 3.Kf4 Nf7! 4.Kf5 Nh6+! 5.Kg5 Nf7+ 6.Kh4 Kg6 3.Kg4 The tablebase shows that 3.Qd7+ is a couple of moves faster 3...Nhf8 4.Qd6 the first edition of the Handbuch gives this move a question mark with no further elaboration. 4...Kf7 5.Qd5+! The only winning move, but it is given a question mark in the final edition, I guess because it is setting up the blunder on the next move. The alternative given 5.Kg5? Ne6+ 6.Kf5 Ne7+ is a draw 5...Kg7 6.Kg5? as shown on the correction link, White is winning 6.Qb7+ Without offering an alternative, the Handbuch gives the text a question mark, because the draw is clear after 6...Nh7+! and White must walk into a queen fork.

As should be apparent, this is a very tricky ending to play correctly from either side, although White is of course running no risk of losing. While there are statistically more drawn positions than with the other minor piece combinations (B+N or B+B) the drawn positions in those endings are stone cold fortresses, whereas with two knights the defender has to keep switching the barrier to keep the White king at bay. Here is an example from the 2014 Central Serbian Team Championship between Nebosja Djokic and Dobrisav Stojanovic. After 69.Kxh3

Usually, having the knights defending each other is a poor defensive formation, but here Black should be holding despite the fact that there are pawns on the board. His formation is very compact with knights blockading the White pawn while defending the e-pawn and the king well centralized. 69...Kf4?! As White doesn't have any good checks, Black could just hold the position with 69...Na5 and 70...Nac6 70.Qd3 Nf3! 70...e4? 71.Qg3+! Kf5 72.Kg2; 70...Ne7 71.Kg2 71.Qd6 Nfd4! 72.Kg2 Ke3 73.Qh6+ Kd3 74.Kf2 Kc4 75.Qc1+ Kd3 76.Qf1+ Kd2? Black needs to stay near his pawn with 76...Ke4 77.Qe1+?White wins with 77.Qc4 as played a few moves later or with 77.Qh3 cutting off the Black king 77...Kd3! 78.Qf1+ Kd2? 79.Qc4 getting it right on the second chance 79...e4 80.Qa2+! Kd3 81.Qa6+?! Kd2 forcing White to repeat the position 82.Qa2+! Kd3 83.Qb1+ Kc4

84.Qxe4? White needs to use his king to help keep the Black forces disorganized 84.Ke3! Kxc5 85.Kxe4 84...Kxc5 85.Ke3 Kc4? The best defensive formation for Black is setting up a barrier with the knights next to each other defended by the king so the way to defend is 85...Nb5! bringing this knight to d6 or c7 86.Qd3+ Kd5 87.Qa6? Ke5? 87...Nf5+ followed by 88...Nd6 sets up Black's ideal defensive position.

88.Qc4? The only winning move here is the hard to find 88.Qa1! The logic is to prevent the maneuver Nf5-d6 88...Nf5+! 89.Kd3 Ncd4? playing one of the knights to e7 gives Black time to regroup 90.Qc5+ Kf4 91.Qc7+ Kf3 92.Qe5 Kg4 93.Ke4 Kg5 94.Qf4+ Kf6 95.Qg4 Kf7 96.Ke5 Ke7 97.Qg8 Kd7 98.Qd5+ Kc7 99.Qc5+ Kd7 100.Kd5 Kd8 101.Qa7

101...Kc8 101...Ne7+ is more resilient 102.Ke4! (102.Kc4? Ne6=) 102...Ne6 103.Qb6+ Kd7 104.Qb5+ Kd8 105.Ke5 102.Qf7 Kd8 103.Kc5 Kc8 104.Qe8+ Kc7 105.Kd5 Kb7 106.Qd7+ Kb6 107.Kc4 Ka6 108.Qc7

Stalemating the Black king, so the knights will have to give up their mutual defense after which one of them will soon be lost. 1-0

3/8/20 - He-Woo, 2019 US Masters

There is an adage about knight endings that they play much like pawn endings. I find this too be a bit too simplistic. It seems to be much harder to give general rules on knight endings because so many positions are precise and depend on what sort of circuit the knights find themselves on.

Several knight endings with limited material have caught my eye over the past several months. The first one is from the 2019 US Masters between two young stars, Anthony He and Christopher Woo. After 47.bxc5

Black is better because his king is better place and his knight is much better placed. The White c-pawn is somewhat weak and Black should be winning material. 47...Nc3+ 48.Ke3 Kd5? Black could win the a-pawn with 48...Nb1 49.a4 (49.c6 Nxa3 50.c7 Nc4+ 51.Ke2 Nb6) 49...a5 50.c6 Nc3! 51.Ng2 (51.c7 Nd5+) 51...Nxa4 52.c7 Nb6 49.c6 Kxc6 50.Kd4! Nb5+ 51.Kxe4! Nxa3 52.Kd4 a5 53.Nf3 h5 54.Kc3 Although this should still hold, I don't like volunarily giving ground with the king. 54.h4 Nb5+ 55.Kc4! Nd6+ 56.Kd4! a4 57.Ne5+ Kb5 58.Kc3 Nf5 59.Ng6 Kc5 60.Nf4 Ng3 61.Kb2 Kb4 62.Ka2 a3 54...Kd5

55.Kb3? The consistent followup to the previous move, but the king turns out to be too far astray here. However, the drawing margin was starting to get pretty thin. 55.Nd2 Ke5 56.Nb3 Nb5+ 57.Kc4! Nd6+ 58.Kc5! a4 59.Nc1! h4 60.Kb4! Kf4 61.Kxa4! Kf3 62.Nd3!

Now, the direct 62...Kg2 63.Nf4+ Kxh2 64.Ng6 h3 65.Nf4 and 66.Nxh3 is an immediate draw, but it takes the Black knight too many moves either cover f4 or to take the h-pawn. 55...Nc4! 56.Ka4 56.Ng5 Nb6 and White isn't getting the a-pawn 56...Ke4! 57.Ng5+ Kf5 58.h4 Ne5! 59.Kxa5 Ng6! 60.Kb4 Nxh4!

61.Nf7 One adage about knight endings that usually holds including this game is that knights struggle against rook pawns. Here, 61.Nh3 also loses 61...Kg4 62.Nf2+ Kf3 63.Nh3 Nf5 64.Kc3 h4 65.Kd2 Nd4! 66.Ke1 Kg3 67.Nf2 (67.Ng5 Kg2! 68.Kd1 Nf3) 67...Kg2! 68.Nd3 Ne6 69.Nf2 Nf4 70.Ng4 h3 71.Ne3+ Kf3 72.Nf1 Nd3+ 73.Kd1 Kf2 74.Nh2 Ne5-+ 61...Nf3 62.Kc3 h4! 63.Kd3 h3 64.Nd6+ Kf4 65.Ne4 Ne5+ 66.Kd4 Ng4 0-1

3/4/20 - BCE-231, Romanovsky-Verlinsky, Moscow 1925

The week's BCE position is again from a game where a knight proved superior to a bishop. Unlike last week, with slightly less material, the bishop should have been able to hold the draw.

I still haven't gotten a copy of Bogoljubow's book on the historic Moscow 1925 tournament, but there is extensive analysis of today's ending by Averbakh in the Encyclopedia of Chess Endings. Outside of Bogoljubow, who won the tournament, Peter Romanovsky was the most successful of the Soviet players finishing tied for 7th in the 21 player field. The loss in this game resulted in Boris Verlinsky finishing with a -1 score, but had the memorable achievement of beating the World Champion, Capablanca, with the Black pieces.

In the knight vs. bishop ending after 51.Nxb4 the pawns are even but Black is doomed to lose one of his doubled f-pawns.

51...Bb7 52.Nd3 Kf6 53.Nxf4 Bc6 The starting position for BCE-231 54.Ke3 Kg5 55.g3 Ba4 Fine gives the variation 55...Kg4 56.Kf2 Be4 57.h3+ Kg5 58.Ke3 Bc6 59.Ne6+ Kf6 60.Nd4 Bg2 61.h4 Ke5 62.Nf3+ Kf6! 63.Kf4 saying that it transposes to a position that occurred in the game, but the bishop never goes to g2 in the game. In any case, Black holds here with 63...Kg6! preventing the advance of the h-pawn which would be followed by Nh4. 56.Nh3+ Kg4 57.Nf2+ Kg5 58.h3 Bc6 59.h4+ Kf6 We will see in many variations below that g6 is a good square for the Black king, preventing h5 and allowing a counterattack with Kh5-g4 in some situations. Still, the move to f6 has logic behind it. The king covers e5 and g5 from White's king invading. It seems like it will be hard for White to check him off of f6 with his knight. However, leaving the h-pawn free to move gives White a spare tempo with h5 which can results in some zugzwang positions. 60.Kf4 Be8 61.Nd1 Bf7 62.Ne3 Be6! Averbakh's analysis begins from this point as ECE Position #973. Defending f5 from the other side with 62...Bg6? allows the White king to penetrate 63.Nd5+! Ke6 64.Kg5 Bh7 65.Ne3 Ke5 66.h5 Ke4 67.Nxf5 Bxf5 68.g4 Bh7 69.Kh6+- 63.Nc2 Bf7 Black can also maintain his bishop on the c8-h3 diagonal 63...Bc8 64.Nd4 Bd7 65.Nb3 Bb5 66.Nc5 Bc6 67.Na6 Bd5 68.Nc7 Bf7 69.Na8 Bd5 70.Nb6 Bf7 (Averbakh considers 70...Bg8? in ECE-974. White wins, but there is one flaw in one of his subvariation 71.Nd7+ Ke6 72.Nc5+ Kf6 73.h5 Bh7 74.Nd7+ Ke7 75.Ne5 Kf6 76.Nf3? (Averbakh's main line is the winning 76.h6) 76...Bg8 77.Nh4 and here 77..Bf7 draws in addition to Averbakh's main line of (77...Be6) 78.h6 Ba2 (instead of Averbakh's 78...Be6? which allows a fork trick 79.h7! Kg7 80.Ng6! Kxh7 81.Nf8+!) 79.Nxf5 Kg6 80.g4 Bb1

and White can't make progress because the double attack on the knight prevents the advance of the g-pawn and the h-pawn drops off if the knight moves. Even if White brings his king around to advance the h-pawn, he gets stuck in the corner 81.Ke5 Bc2 82.Ke6 Bb1 83.Ke7 Bc2 84.Kf8 Bb1 85.Kg8 Ba2+ 86.Kh8 Bb1 87.h7 Bxf5 88.gxf5+ Kf7! 89.f6 Kf8! 90.f7 Kxf7!) 71.Nd7+ Ke6! 72.Ne5 Be8! 73.Nc4 Kf6! 74.Ne3 Bd7! 75.Nd5+ Black holds even losing the f5 pawn with 75...Kg6! (instead of Averbakh's 75...Ke6?) 76.Ke5 Kh6! 77.Nb6 Bc6! 78.Kxf5 Kh5!

The fact that this position is drawn is the key to Black's defense. However, he must be attentive to be able to switch his bishop between diagonals that the White knight can't block. For instance, 79.Nc4 Bd7+ 80.Kf4 Be6 (80...Bc8? 81.Nd6 Ba6 82.g4+ Kxh4 83.Nf5+ Kh3 84.g5 Bc4 85.g6! Bb3 86.g7 Bc4 87.Kg5 Bb3 88.Nh6 Bc4 89.Kf6 Bb3 90.Nf7+- 81.Nd6 Bb3 82.Nf5 (82.g4+ Kxh4! 83.Nf5+ Kh3 84.g5 Bf7! is the difference) 82...Bd1! 83.Nd4 Bg4= 64.Na3 Bd5 65.Nb5 Be6 One of Averbakh's lines here has a couple of interesting points 65...Ba2 66.Nd6 Bb1? (66...Be6! 67.h5 Bd7! is a mutual zugzwang position that occurs in the game.

Here, it is White to move and the position is drawn as Black will have Be8 on a knight move, Kg5 on a king retreat or Kg6 if the pawn advances.) 67.Ne8+? (White needs to precede this with the pawn advance 67.h5! Bc2 68.Ne8+! Kf7 69.Nc7 Kf6 70.Nd5+ Ke6 71.h6+-) 67...Kf7 68.Nc7 Kg6! 69.Nd5 Bc2 70.Ke5 Kh5 71.Nf4+ Kh6! 72.Kf6 Bb1 This is ECE-975 73.Ne6 Kh5! 74.Kg7 Kg4! 75.Nf4 Kxg3! 76.h5

Averbakh ends the variation here with a +- evaluation, but Black draws after White promotes because his bishop can deflect the White queen, allowing a drawn f-pawn vs. queen ending 76...Kxf4! 77.h6! Kg3! 78.h7! f4! 79.h8Q f3! 80.Qh1 f2! 81.Qf1 (81.Qxb1 Kg2! and the White king is too far away) 81...Bd3 82.Qxd3 Kg2!= 66.Nd6 Bd7? The BCE correction is that Black draws with 66...Kg6! when the best White can do is reach a drawn position with 2 extra pawns similar to the previous diagram 67.Ke5 Bd7! 68.Nc4 Bc8 (68...Kh5 69.Ne3 Kg6 70.Nd5 Kh6! 71.Ne7 Kh5 72.Nxf5 Ba4!) 69.Nb6 Bb7 70.Nd5 Kh5! 71.Ne3 Bc6! 72.Kxf5 Bd7+ 73.Kf4 Ba4! 67.h5! The mutual zugzwang position we saw earlier. Now it is Black to move and he must allow Ne8+ or Nxf5 67...Be6 68.Ne8+! Kf7 69.Nc7 Bc8 70.Kg5 Bd7 71.Nd5 Kg7 72.Ne3 Kh7 73.Nxf5 1-0