Shakmaty Bereolos - The Official Chess Site of Peter Bereolos

4/29/20 - BCE-368

We saw in the Kasparov-Karpov ending that the defending side has drawing chances with 3 vs. 3 on one and a passed a-pawn even when the attacking rook was behind the pawn. When the rook is in front of the pawn, winning chances are even more difficult. BCE-368 is an example of this. Generally, with a rook in front of the a-pawn, White needs some weaknesses in the Black position in order to win. At first sight, it looks like f5 and h5 provide those weaknesses. However, it turns out that the f5 pawn is largely irrelevant and h5 can be defended. White does have a winning try that was not mentioned by either Fine or Benko. In the first round game of the 1996 Pardubice Open the following position arose in the game Pcola-Hybl after 50...Kg7

51.h4!? this aims to prevent the ...h4 resource that is seen in the correction link 51...gxh3+? 52.Kxh3 f4 Black may have thought he was leaving White with a and g pawns, which would be a draw, but White dashed that hope with 53.f3! fxg3 54.Kxg3! and White wins by pushing the f pawn as we saw in the Shamkovich-Liberzon game. 1-0

Black does better not to capture en passant. In the Albania-Rwanda match at the 2008 Olympiad, the diagram position above arose in the game Guxho-Ngendo after 39...Ra3

40.h4!? Ra2 41.Kf1 Kh7 42.Ke1 Kg7 43.Kd1 Kh7 44.Kc1 Kg7 45.Kb1 Ra5 46.Kb2 Kh7 47.Kb3 Kg7 48.Kb4 Ra1 49.Kc5 Kh7 50.Kd6 Kg7 51.Ke5

51...Ra5+? The simplest defense was 51...Ra4 preventing Kf4. Losing the f5 pawn is not of importance here as after 52.Kxf5 Ra5+! White can't make progress despite his two extra pawns. Black checks on a5 if the White king goes after the kingside and along the b-file if he goes to b6 or b7. White can get the opposition in the pawn ending, but even this is drawn. 53.Ke6 Ra6+ 54.Kd5 Ra2 55.Kd6 Ra6+ 56.Kc5 Ra2 57.Kb6 Rb2+ 58.Kc6 Rc2+ 59.Kd6 Ra2 60.Rc8 (on 60.Re8 Black checks the king away before taking the pawn 60...Ra6+!) 60...Rxa7 61.Rc7+ Rxc7 62.Kxc7 Kf6 63.Kd6 Kf5 64.Kd5 Kf6 65.Kd6 Kf5 and the attempt to outflank loses after 66.Ke7? Ke4!-+; White can also try 52.Ke6!?, but 52...Ra6+! is similar to the previous line 52.Kf4 Now Black is in zugzwang. He needs to keep the rook on a5 to meet Kg5 with f4+, so he makes the only move with his king that doesn't lose immediately. 52...Kh7 He could also have tried the trap 52...Ra4+!? when White wins with 53.Kg5! but not (53.Kxf5? Ra5+! with a draw as we saw above) 53.Rf8 pointing out the problem with Black's king move, now White harvests the Black kingside 53...Rxa7 54.Kg5! Kg7 55.Rxf5 Ra6 56.Kxh5 Ra4 57.Rf4 1-0

4/25/20 - TCEC Season 17 Superfinal, Games 3-4: Sicilian Scheveningen

There has been one high level chess event that has been completed despite the current lockdown. The Super Final of the 17th season of Top Chess Engine Championship (TCEC) between Stockfish and Leela Chess Zero (Lc0) completed this week with Lc0 taking the 100-game match +17 =71 -12. I hadn't really paid much attention to engine chess tournaments before although I of course had looked at the Alpha Zero versus Stockfish games. However, I have heard some people comment that these games should be studied because these are now the best players on the planet. The estimates of the ratings of Stockfish and Lc0 are in the upper 3400's.

One of the criticisms of the first Alpha Zero games against Stockfish was that they were played under conditions that were not optimal for Stockfish. The TCEC is a more neutral competition where all the engines are running on similar hardware. This is a bit more difficult now with the neural net programs like Lc0, which are GPU based versus traditional engines like Stockfish which run on the CPU. The TCEC tried to deal with that issue this season a bit by separating the preliminary rounds into CPU and GPU divisions.

One element that is unique to TCEC is that the games do not start from the opening position. Instead, 50 different opening positions are selected and each side gets a chance with each color. This makes it an interesting crucible for opening theory. I'm going to do some surveys of some of the openings from the match and see how they compare with human practice.

The first decisive games came early, in games 3 and 4, where White won both games from a sideline in the Keres Attack of the Sicilian Scheveningen. 1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4 Nf6 In the 14th game of their second match , Kasparov tried to avoid the Keres Attack against Karpov with the move order 4...Nc6 5.Nc3 d6, but Karpov all the same threw the g-pawn forward with 6.g4 5.Nc3 d6 6.g4 An early thrust of the g-pawn is common in modern chess, but it was quite sensational when Keres introduced it against Bogoljubow in 1943. This move has really put players off of playing the Scheveningen variation. 6...a6 Should this move, which was Petrosian's frequent choice against the Keres Attack, be given a dubious mark? Since 2010, in games where both players were rated at least 2400, White has scored a commanding +16 =6 -4. The most common move is 6...h6 which Kasparov played against Karpov in the first game of their first match. That was the only time Kasparov played the traditional Scheveningen in a World Championship match. 7.g5 Nfd7

This was the starting point for the TCEC match. 8.Be3 The most common move, this was Lc0's choice. Negi recommends 8.Bg2 trying to discourage ...b5. Stockfish went for the more restrained 8.a3 when Lc0 decided to do without ....b5 and developed with 8...Nc6. 8...b5 9.a3 Bb7 10.f4 The only high level game I found with this move was a rapid game where Amonatov deafeated Artemiev. The text move makes quite a bit of sense as attacking the Black structure with f5 is thematic in these positions. The most common move is 10.h4 when Negi suggested that Black is fine after 10...Nc6 11.Nxc6 Bxc6 12.Qd4 Rb8 intending to trade queens with ...Qb6. This got played in the second game of a recent Banter Blitz match between Carlsen and Grandelius. Neither player mentions Negi's book, so I think they were on their own. Carlsen mentions that he is a tempo up on the h3 Najdorf which we can see by looking at a line like 1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 a6 6.h3 e6 7.g4 b5 8.g5 Nfd7 9.a3 Bb7 10.Be3 Nc6 11.h4 I think this offhand comment by Magnus pinpoints why ...a6 against the Keres Attack is out of favor. Who really wants to give White an extra tempo in the Najdorf? Grandelius mentions the idea of trading queens, but after 13.0-0-0 he changed his mind and went 13...a5 perhaps out of respect for the World Champion's endgame prowess. Unfortunately, Carlsen blundered a few moves later, so the game lost theoretical interest, although he still hung on to draw. On a historical note, the earliest game in my database in this line is Fischer-Najdorf from the 1960 Olympiad. There Bobby chose 10.Qd2. 10...Nc6 11.Rg1!?

The novelty, Amonatov captured on c6. At first this move might seem a bit obscure, but it takes the rook off of the diagonal from Bb7, gives extra protection to the g-pawn in preparation of f5, and also allows for a rook lift Rg3 which can give extra defense to the queenside. Stockfish itself also chooses 11.Rg1 giving White a large advantage, over +1.4. I find this interesting because presumably Stockfish was playing the moves it thought best for Black after 8.Be3, but Stockfish chose 8.a3 when it had White. A couple of versions of Stockfish on Chessbase's Let's Check rate that move around +1.0.

I do have one experience in this line. At the start of our sophomore year Billy Colias and I played a 5-game match to determine who would start the year as first board for the Munster High School team, which Billy won 3-2. In the third game, I abandoned my normal 1...e5 and played this variation of the Scheveningen. From the first diagram, the game continued 8.Be3 Nc6 9.h4 Qc7 10.Qd2 Nde5 11.f4?! Nxd4 White can't recapture either knight beause of the fork on f3. Billy found a way out with 12.0-0-0 and Black can't hang on to his extra piece. I grabbed a pawn and exposed his king with 12...Nxc2 and eventually lost a long game. Trying to hold the piece with 12...Nef3 13.Qf2 e5 doesn't work as after 14.Nd5 Qa5 15.Kb1 White is going to play c3 and one of the knights falls. A reasonable alternative is 12...Ndc6 13.fxe5 Nxe5

I think the reason I don't often play the Sicilian is that I would rate this position as clearly better for Black. He has an extra pawn, no weaknesses, and a beautiful outpost for his knight on e5. He might be able to grab the bishop pair with ...Nc4 and White has an isolated e-pawn. For compensation White has a small lead in development and more space on the kingside. Yet Stockfish gives the Black advantage as only -0.29 after 14.h5

4/24/20 - Andrews-Bereolos, 1996 Music City Grand Prix

One of my earliest games against Todd Andrews featured an ending with 3 connected passed pawns versus a rook. After 31.g3

Black is a pawn down, but his active rook gives some compensation 31...h4? Trying to attack White's kingside, but there is a flaw. It was better to activate the king with 31...Kb7 32.Rd1? 32.gxh4 Bxf4 White has a surprising way to guard h2 with 33.Be5 Now, the rook ending doesn't look too good for Black 33...Bxe5 34.Rxe5 Rxb2 35.h5 Rb4 36.Rf5, but the passed h-pawn is going to cause Black problems even if he keeps the bishops on the board 32...h3 Even though Black temporarily loses another pawn with this move, the Black pawn on h3 is very useful in confining the White king to the back rank and Black's problems are behind him. 33.Bxb6 axb6 34.Rxd6! Rg2+ this thematic idea appears several times in the ending. Black poses White the choice of protecting or abandoning his h-pawn 35.Kf1 35.Kh1?! doesn't make a lot of sense here as 35...Rxb2 36.Kg1 Rg2+ repeats the position except that White has dropped his b-pawn. 35...Rxb2 36.Rd8+?! better is 36.Rf6 as the text just activates the Black king 36...Kb7 37.Rd7+ Kc6 38.Rxf7 Rxa2 39.Kg1?! White is really skating on thin ice after this move. There is not much problem making a draw with 39.f5 Rxh2 40.f6 Rg2 41.Rg7 Kd6 42.Rb7 Rxg3 43.Rxb6+ Ke5 44.f7 Rf3+ 45.Kg1 Rxf7 46.Kh2= 39...Rg2+!? With the idea of putting the rook on the other side of the b-pawn before advancing it. 40.Kh1 White doesn't want to admit his previous move was wrong, but moving the king to the f-file would prevent the black rook from stopping the f-pawn so the pawn race should end in a draw 40.Kf1 Rxh2 41.Rh7 b5 42.Rh6+ Kc5 43.Rh5+ Kc4 44.f5 b4 45.f6 b3 46.Rh4+ Kc3 47.f7 b2 48.Rb4! Kxb4 49.f8Q+! and White will give perpetual check 40...Rd2 41.Kg1! b5

42.Rh7? The only way to draw is 42.Rf8! to get the rook to b8 before the b-pawn becomes too far advanced Black can again try 42...Rg2+!? 43.Kf1! (43.Kh1? b4 44.f5 b3 45.f6 b2 46.Rb8 Rd2 47.Kg1 Rd1+! 48.Kf2 b1Q 49.Rxb1 Rxb1! 50.f7 Rb8 -+) 43...Rxh2 44.g4 (44.f5? Rh1+! 45.Kf2 Ra1 46.Rh8 h2! 47.Rxh2 Ra2+! 48.Kg1 Rxh2 49.Kxh2 b4-+) 44...b4 (now on 44...Rh1+ 45.Kf2 Ra1 White has 46.Kg3) 45.Rb8 Kc5 46.f5 Kc4 47.f6 Rh1+ 48.Kf2 h2! 49.Kg2 Rf1 50.Kxh2 Rxf6=; 42...b4! 43.Rxh3 b3! 44.g4 b2! 45.Rb3 Kd5! The White pawns get far enough advanced to draw after 45...Kc5? 46.f5 Rd1+ 47.Kf2 b1Q 48.Rxb1! Rxb1! 49.f6= 46.Rb4 Now on 46.f5 the Black king is close enough to slow the pawn's advance 46...Ke5! 47.Rb5+ Kf4! (but not 47...Kf6? 48.h4 Rd1+ 49.Kf2 b1Q because of the zwischenzug 50.g5+!) 48.Rb4+ (48.f6 Rd1+! 49.Kf2 b1Q 50.Rxb1 Rxb1! 51.f7 Rb8) 48...Kg5-+ 46...Rd1+ 47.Kf2 b1Q 48.Rxb1 Rxb1!

The White pawns need to be one rank further advanced in order to draw 49.Kg3 Ke4 50.f5 Rb6 51.Kh4 Kf4 52.Kh5 Ra6 53.h4 Rb6 54.f6 Rxf6! 55.g5 Ra6 56.g6 Kf5 57.g7 Rg6 0-1

4/22/20 - BCE-301, Euwe-Capablanca, Karlsbad 1929

This week's BCE position is another game from Karlsbad 1929. This time two World Champions are featured, but neither held the title at the time of this game. As mentioned last week Capablanca was half a point short of Nimzowitsch's pace. He lost 2 games in the event after going +30 -1 =36 from 1926-1928. Euwe finished tied for 5th with a +3 score. Their game is actually featured in two BCE examples. BCE-398 starts with the position after 39.Rg6

39...Kc5 40.Rxh6 Kd4 41.Kf3 a5 42.Kf4 a4 43.Rh5? Fine's main line of 43.Rh3 wins as White saves many tempi versus the game line. 43...Ra6 44.Rd5+ Kc3 45.Rd1 a3 46.Kf5 a2 47.f4 a1Q 48.Rxa1 Rxa1 49.Kxf6 The starting position of BCE-301 49...Ra6+ 50.Kf5 Kd4 51.e5 Kd5 52.g3 Ra8 53.Kf6 Ra6+ 54.Kf5 Ra8 1/2-1/2

4/18/20 - Palatnik-Bereolos, Chattanooga 1997

I've uploaded my annotations to my 1997 game against Sam Palatnik from the final round of the Living Legend tournament. This was the 2nd year that the Chattanooga Chess Club held an event to honor their senior member, Rea Hayes. This was during the time that Palatnik was the Tennessee GM in residence and it was not unusual for him to participate in a 1-day event. Despite getting a reasonable position from the opening, I don't think I gave him too much trouble. This game is a good lesson in good pieces versus bad pieces.

4/17/20 - Shamkovich-Liberzon, Moscow Jubilee 1967

One of the big steps forward in chess information was the pubication of Chess Informant beginning in 1966. It's symbolic system allowed players around the world to read annotation by the world's best players without language barriers. The 5th volume introduced the Combinations and Endgames section, and today I'm going to look at one of the rook endings from that issue.

The Moscow Jubilee tournament in October 1967 was part of the celebration of the 50th anniversary of the Russian Revolution. Unlike the very strong international event held in Moscow earlier in the year, this one was an all-Soviet affair. It was pretty strong with 8 of the 12 participants holdig the GM title. 3 of the other 4 went on to become GMs, the sole exception being Gennady Ageichenko. However, Ageichenko is inspirational in that he finally made the IM title at age 69! Lev Polugaevsky and former World Champion Vassily Smyslov tied for first with undefeated +5 scores

The opponents in today's game were two of the first grandmasters to emigrate from the Soviet Union. Vladimir Liberzon became Israel's first GM and played top board for them in several Olympiads. Leonid Shamkovich eventually settled in the US and was a frequent participant in the US Championship. Neither fared great in the Jubilee tournament. Shamkovich had -1, while Liberzon withdrew with illness after scoring 2.5/8. However, their encounter did produce a rook ending that is worth study. Judovic gave light notes in Informant #5. Keres analyzed the game in depth in the March 1968 issue of Chess Life. Finally, Müller and Konoval analyzed it using tablebases. So there wouldn't seem to be much to add, but there is a critical positions right at the start of the rook ending. White had an extra pawn after 46.a4

46...Bd4? After Keres' suggestion of 46...Bc3 Black is probably objectively holding, but White can still press for a win. Instead, Black tries to take refuge in a rook ending, but it is losing for Black. 47.Bxd4 Rxd4 Without the h-pawn, this is a textbook win for White by advancing the a-pawn to the seventh then pushing the f-pawn. The question is if the h-pawn makes a difference. 48.a5! The analysis in Informant and by Müller and Konoval did not start until after this critical move. White can't start with the rook move 48.Ra8? Rg4+! 49.Kf3 (49.Kh3 Rf4 50.Kg3 Rg4+; 49.Kf1 h4! 50.f3 Rf4 51.Kf2 h3 52.a5 h2 53.Kg2 h1Q+ 54.Kxh1 Rxf3!=) 49...Rg1! 50.a5 (50.Rh8 Ra1) 50...h4! 51.a6 h3! 52.a7 h2! 53.Rh8 (53.Rf8+!? Kg7 54.Rg8+!? Kf7) 53...Ra1! 54.Rxh2 Kg7; Preventing the rook check with 48.f3? doesn't help either 48...h4 49.a5 Ra4 50.Ra8 Ra2+ 51.Kh3 Ra3 52.Kg4 Ra4+ 53.f4 h3 48...Rg4+ Keres shows that Black is a tempo short if he tries to trade the kingside pawns 48...h4 49.Ra8 Rd2 50.a6 h3+ 51.Kxh3 Rxf2 52.a7 Ra2 53.Rh8! 49.Kf3 Ra4 49...Rg1 50.Ra7+! Kg6 51.a6 h4 52.Ra8! Ra1 53.a7 Kg7 54.Kg4+- 50.Ra8 Kg6 51.a6 Kg5 52.Ke3 Judovic's variation is the simplest way to win 52.a7 Kh4 53.Kg2 Ra2 and now a trianglualtion to put Black on move 54.Kg1 Ra1+ (Keres also shows that the tougher defense with 54...Kh3 also fails 55.f4! h4 56.f5! Ra5 57.f6! Ra6 58.f7! Rg6+ 59.Kh1! Ra6 60.Rg8+-) 55.Kh2 Ra2 56.Kg2 Ra1 57.f4 and wins in similar fashion to the game. 52...Ra3+ 52...Kh4 53.a7! Kh3 54.f4! Ra3+ 55.Kd4 Ra4+ 56.Kc5 Kg4 57.f5! Kxf5 58.Rf8+! Kg4 59.a8Q Rxa8 60.Rxa8! h4 61.Kd4! h3 62.Ke3! h2 63.Rg8+ Kh3 64.Kf2! h1N+ 65.Kf3! and Black loses his knight 53.Ke4? 53.Ke2 Kh6 54.a7 Kh7 55.f4 h4 56.f5 h3 57.Kf2! and the White king gets in front of the h-pawn

53...Ra4+? With the White king cut off along the third rank, Black can retreat his king and use his h-pawn for counterplay 53...Kh6! I'll flaunt the Nunn Convention here and award this move an exclam (53...Kg6 also holds the draw). I think too often these days and declare moves blunders based on the engine. Over the board, I think it is incredibly hard to find this king retreat, especially since only two moves earlier the king advanced with the idea of sheltering in front of the h-pawn. 54.a7 Kh7! (not 54...Kg7? which allows the White pawn to advance to f6 with check 55.f4! h4 56.f5! h3 57.f6+! Kh7 58.f7! h2 59.Rh8+! Kxh8 60.f8Q+!) 55.f4 h4! 56.f5 h3! 57.f6? h2! 54.Kf3? Blocking the f-pawn costs White precious time. He should have repeated the postion with 54.Ke3; Keres' suggestion of 54.Ke5? fails for the same reason we saw on the previous move 54...Kh6! (instead of 54...Ra5+? but it is not surprising that he missed this as he let move 53 pass with no comments. ) 54...Kh4! 55.Ke3 Ra3+ 56.Ke4 Ra4+ 57.Kf5 Ra5+ 58.Ke6

58...Kg4? Judovic gives 58...Ra4? an exclam with no further analysis, but Keres shows that White wins with 59.a7! Neither annotator shows the drawing line 58...Kh3! 59.a7 h4 60.f4 Kg4 61.f5 (61.Rg8+ Kxf4 62.a8Q Rxa8! 63.Rxa8! h3=) 61...Ra6+! 62.Kd5 Kxf5 63.Rf8+ Kg4 64.a8Q Rxa8! 65.Rxa8! h3 66.Ke4 h2= 59.a7! Ra6+ 60.Ke5 Keres shows one last trap White could fall into with 60.Ke7? Kh3 61.f4 Kg4! and the White king is too far afield to help in the coming rook vs. pawn ending. 60...Ra5+ 61.Ke4 Ra4+ 62.Ke3 Ra3+ 63.Ke2 Ra2+ 64.Kf1 Kh3 on 64...Ra1+ White can hide his king with 65.Kg2 which would not have been possible if Black had played 58...Kh3 65.f4! Ra5 66.Ke2 Ra2+ 67.Kd3 Ra3+ 68.Kc4 Kg4 69.Rg8+ Kxf4 70.a8Q Rxa8 71.Rxa8! h4 72.Kd3 h3 73.Ke2 1-0

4/15/20 - BCE-403a, Bogoljubow-Maroczy, Karlsbad 1929

This week's BCE position is from the 1929 Karlsbad tournament, which was Nimzowitsch's greatest triumph with 15/21 half a point clear of Capablanca and Spielmann. The two big names involved in this game finished in the middle of the pack. Bogoljubow was 8th at +2, while Maroczy was =12th at -1.

I have Nimzowitsch's book on the event, but he did not cover all the games and this one was omitted. Other sources are still a few years from being in the public domain, so I didn't have any other analysis to compare my notes with. There was an interesting moment right at the transition into the rook ending after 32...Rd8

33.Rb1 33.Ke3!? exd5 (White wins the pawn ending after 33...Rxd5? 34.Rxd5 exd5 35.Kd4 Ke6 36.h4!) 34.Kd4 and it seems that the black pawn is much more vulnerable on d5 than on e6. 33...Rxd5+ 34.Ke3 Rd7 35.Ke4 Kf6 36.Rb3 Rc7 37.Rc3 Rc6 38.Kd4 b6 A somewhat instructive moment. Black dissolves White's weak pawn in order to activate his rook. It doesn't look like White has much of a plan, but look what can happen if Black tries to just sit passively 38...Rc7 39.a4 Rc6 40.a5 a6 (40...Rc7 41.a6 bxa6 42.c6) 41.Rb3 Rc7 42.Rb6 and White will press further with c6. 39.a4 bxc5+ 40.Rxc5 Rd6+ 41.Kc3 Rd1 42.Rb5 42.Ra5 Rf1 43.Rxa7 Rxf2 transposes to the game 42...Ra1 43.Ra5 The starting position of BCE-403a 43...Ra2 44.Rxa7 Rxf2 45.Rh7 h5 46.a5 Fine's main line with 46.g4 is a simpler way to draw. 46...Rxg2 47.a6 Ra2 48.a7 Kg5 49.Re7 Kh4 50.Kb3 Ra1 51.Rxe6? 51.Rc7 is the subject of the BCE correction 51...Rxa7! 52.Rxg6 Kxh3! 53.Kc2 Rd7! 54.Rg5 h4 55.Rg1 Rd4 56.Kc3 Rg4 57.Ra1 Kg2 58.Kd3 h3 0-1

4/11/20 - Wagner-Bereolos, 1997 Kings Island Open

Mark Wagner is another one of my once a decade opponents. I've previously discussed the endings I played against him in 2007 and 2019. Our first meeting, in the second round of the 1997 Kings Island Open reached a materially even knight ending after 40...Nxe8

Black is very slightly better here because he will be able to create an outside passed pawn on the queenside with ...d5. Still, this isn't much. 41.h5 gxh5 42.Nxh5 d5 43.cxd5 Nc7!? provoking the d-pawn forward in the hopes that it will become weak. Another try is 43...Nd6 44.Ng3 Nxb5 45.Nf5 Nc3 46.Nxh6+ with similar play 44.d6 Nxb5 45.d7 Ke7 46.Nf6 Nc7 47.Ng8+? 47.f4 b5 48.Ne4 Kxd7 49.Nxc5+ Kc6 50.Nb3 Kd5 is equal 47...Kxd7 48.Nxh6 Compared to the variation after 43...Nd6, Black is well ahead since he has already collected the d-pawn 8...Ne6 49.Kg3 b5 50.Nf7 Nc7? Black should win with 50...Kc6 with the idea of pushing the c-pawn. Black's pawns are much faster than Whites: 51.f4 c4 52.dxc4 bxc4 53.f5 Nc5 54.Kf3 c3 55.Ke2 d3+ 56.Kd1 c2+ 57.Kc1 Nb3+ 51.g5 Ke6 52.Nd8+ Kf5 53.Nf7 b4 54.f4? 54.Nd6+ likely leads to a draw as 54...Kxg5 is met by 55.Ne4+ and 56.Nxc5 54...Nd5? Black should take control of d6 with 54...Nb5 when he should be winning 55.Nd6+ instead Black has to waste a tempo because of the check 55...Kg6 56.Nc4 b3 57.Kf3 Ne3

58.Ke4? A complete breakdown. White could draw with either knight retreat 58.Nd2 b2 59.Ke4 Nf1 60.Nb1 Ng3+ 61.Kd5 Ne2 62.Kxc5 Nc3 63.Nd2= or 58.Nb2 Kf5 59.Ke2 Nd5 60.Kd2 Nxf4 61.Na4 Ne6 62.Kc1 Kxg5 63.Kb2 Kf4 64.Kxb3 Ke3 65.Kc4= Kd2 66.Nxc5 Nxc5 67.Kxd4! 58...Nxc4 59.dxc4 b2 60.Kd5 b1Q 61.Kxc5 Qf5+ 62.Kxd4 Qxf4+ 63.Kd5 Qxg5+ 64.Kd6 Qd8+ 65.Kc6 Qc8+ 0-1

4/10/20 - Shen-Furline, 2019 CCCSA Fall IM

Another endgame that was key in Christopher Shen's IM norm performance in the Charlotte Chess Center's Fall IM tournament came in the very first round against Jacob Furfine. Shen appeared well on his way to victory with an extra exchange after 48.fxe3

It seems clear that Black does not have sufficient compensation. All the Black pawns are weak and the White pawns are safely on the opposite color of the Black bishop. The position is open enough that White should be able to penetrate with his rook and/or king. Black's only trump is the passed b-pawn. Having said all that it would seem that White's win should just be a matter of technique, but after analyzing it quite a bit, the win is not at all easy. 48...Bb5 Black uses some tactical nuances to activate his bishop before pushing his pawn. It turns out that the White pieces are not on perfect squares. 49.Rf2 49.Rf5?! only allows Black to activate his king. 49...Kg6 50.Rf2 (50.Rxd5?? Bc6 or 50.e4? Bd3 are the tactical points) 50...Bc4 51.Rb2 b5 52.Kg3 Kf5 53.Kf3 Ke6 54.Kf4 Kd6; Sending the king over to stop the b-pawn doesn't seem to help much either 49.Kf2 Bc4 50.Ke1 b5 51.Kd2 b4 52.Kc2 Ke6 and the White rook is stuck preventing penetration by the Black king via f5. 49...Bc4 50.Rb2 b5 51.Kg3 Ke6 52.Kf4 Bd3 53.Rb3 Bc4 54.Ra3 Kd6 55.Kf5 55.Ra6+ Kc7 (55...Ke7 56.e4 seems to be easier as the Black king is cut off) 56.Rxf6 b4 57.Ke5 b3 58.Rf2 Kc6 59.e4 dxe4 60.Kxe4 Kb5 61.Ke5 Kb4 62.d5 Kc3 63.Rf3+ Bd3 64.d6 b2 65.d7 b1Q 66.d8Q and the engine claims a win for White, but this could be a quite difficult position to win in practice. This is a long and difficult variation, but creating a passed d-pawn with e4 seems the only way to break down Black's defenses 55...Kc6 56.Kxf6? Now the Black king gets too active. There was still time to return to the previous variation with 56.Ra6+ Kb7 57.Rxf6 b4 58.Ke5 b3 59.Rf2 56...b4 57.Ra1 Bd3 58.Ra4 Kb5! 59.Ra8 b3 60.Rb8+ Kc4 61.Ke5 Be4 62.Rc8+ Kd3 63.Rb8! Kc2 64.Rc8+ Kb1 65.Rb8? 65.Kf4 b2 66.Kg5! Bf3 67.Rb8 Kc2 transposes to the game 65...b2! 66.Kf4

66...Kc2? Black is winning with 66...Kc1! 67.Kg5 (67.Rxb2 Kxb2! 68.Kg5 Kc2 69.Kxh5 Kd3) 67...b1Q 68.Rxb1+ Bxb1! when is king is one square closer to e3 than in the the next variation. 67.Kg5! Bf3 Now if Black promotes, the White king returns in time to protect e3 67...b1Q 68.Rxb1! Kxb1! 69.Kxh5! Kc2 70.Kg5 Kd3 71.Kf4! 68.Rc8+ Kd2 69.Rb8! Kc2 70.Rc8+ Kb1 71.Rb8 Be2 72.Rb7 Kc2 73.Kf4 Bd3 74.Kg5 Be2 75.Kf4 Bc4 76.Rxb2+! Kxb2! 77.e4! dxe4 78.Kxe4! Kc3 79.d5! Be2 80.d6 Bg4 81.Kf4 Kd4 82.Kg5! Ke5 83.d7! 1/2-1/2

4/8/20 - BCE-396b, Steinitz, 1885

The first World Champion, Wilhelm Steinitz, also composed a few studies. Today's BCE position was first published in 1885 in The International Chess Magazine, which Steinitz editied. I was unable to find a digitized version of the 1885 volume. Fine's analysis in BCE matches that of Berger in position 267 of Theorie und Praxis der Endspiele, but I can't tell if that Steinitz' solution or if Berger added elaboration.

The position is somewhat similar to Tarrasch-Johner except that here Black captures the a-pawn and leaves White with a kingside pawn which can be stopped. In that game, Black collected all the kingside pawns, but White should have won with his a-pawn.

4/1/20 - BCE-185b, Duras, 1906

Like his contemporary Reti, the Czech grandmaster Oldrich Duras excelled at both over the board play and at chess composition. Over the board, Sonas has him peaking at #4 in the world in 1909 behind only Lakser, Rubinstein, and Capablanca. He won a special prize for the study that is the subject of today's BCE position. Fine cites it as 1906-07, which is strange for a study. I'm going with just 1906, the date that is shown in both the van der Heijden database as well as Southerland and Lommer's 1234 Modern End-Game Studies, which Fine cites as one of his sources. Both of those list the study as being published in Bohemia, but I was unable to find digitized copies of that periodical online to confirm the date. Neither of those sources give the mistake in BCE, so that may have been an addition by Fine. The study is very deep and beautiful, yet clearly of practical value as well. I'll just give the main line

1.Ba3! Kc4 2.Be7! f3 3.Bd8! Bxh2 4.Bb6! Kb5 4...g4 is the subject of the BCE correction 5.a6! g4 6.Bf2! Bc7 7.b8Q+ Bxb8+ 8.Kb7! Ka5 9.Bh4! Kb5 10.Be1 g3 11.Bxg3! Bxg3 12.a7! f2 13.a8Q! f1Q 14.Qa6+! Kc5 15.Qxf1! 1-0