Shakmaty Bereolos - The Official Chess Site of Peter Bereolos

6/26/20 - TCEC Season 17 Superfinal, Games 13-14, Modern Benoni-Classical Main Line

My plan of looking at openings from the TCEC Superfinal has hit a bit of a hiccup in that I am not even through 1/6th of the openings played and already the Season 18 Superfinal has begun (again with Lc0 and Stockfish squaring off). Nevertheless, I think I am going to persist with the Season 17 openings and try to finish that off.

While the Taimanov and Modern Main Line get all the attention these days, the Classical Main Line is still of quite a bit of importance if for no other reason that many Black players only play the Benoni via move orders that avoid those two dangerous tries. The Classical Main Line was featured in Games 13 and 14. Although Lc0 extended its lead in game 14, I found Stockfish's play in the opening a bit strange. Lc0's defense in game 13 probably won't inspire many backers, but is of more importance theoretically.

1.d4 Nf6 2.c4 c5 3.d5 e6 4.Nc3 exd5 5.cxd5 d6 6.Nf3 g6 7.Nd2 This move orderavoids Black's options with ...Bg4. 7...Bg7 8.e4 0-0 9.Be2

The starting position for this set of games. 9...Na6 Stockfish chose 9...Nbd7 which can lead to similar positions if Black carries out the maneuver ...Ne8-c7, which Stockfish attempted. However, it never completed that maneuver nor did it get in ...Re8, which is almost always a part of Black's plan. 10.0-0 Nc7 11.a4 Re8 12.f3 b6 Petrov choses 12...Nd7 as his main line in his GM Repertoire book calling it a modern line. Indeed, it has been played by Gashimov, Topalov, and Ivanchuk in this century. Those names are usually pretty good endorsements, but the move dates to the 1950s when it was played by Tal. There is also a Fischer game from 1968 with this move. 13.Nc4 Ba6 14.Bg5 h6 Watson said he didn't find a way for Black to equalize after this move and Petrov gives it a ?! mark. Both authors prefer 14...Qd7 15.Be3 Bxc4 16.Bxc4 a6 17.Qd2 Kh7 18.Rab1 Rb8 19.b4 b5 20.axb5 sometimes White does not insert this exchange and continues 20.Be2 20...axb5 21.Be2 Nd7

Lc0's novelty. Previously 21...c4 was played. I agree with Petrov's description of this type of position (he didn't include the exchange on b5) as passive and generally unappealing. Lc0's move is much more dynamic and more in the spirit of the Benoni. Black sacrifices a pawn, but takes over the dark squares. This probably isn't enough to interest Black players in the line, but is a good example of resources the engines find he positions where humans might slowly get ground down. Stockfish took the pawn 22.bxc5 Nxc5 23.Nxb5 Nxb5 24.Bxb5 Re7 but then didn't demonstrate any convincing plan against Black's dark-square grip, trading all the way down to an ending with the light-square bishop against the knight, which offered no chances despite the extra pawn.

6/24/20 - BCE-485, Kling, Horwitz, Campbell, Healey, and Zytogorski

BCE-485 is an interesting duel between rook and bishop. BCE and several other sources I found cite the position as coming from The Chess Player's Chronicle, 1856. However, the position shown there is

with the caption Black has the move; White cannot win. This Chess study is founded on a position occuring in a game of Mr. Zytogorski. The full game seems to be lost to history.

The earliest publication of the BCE position that I found was in 1890 in Berger, which is likely where Fine got the position. Berger also references the position as coming from The Chess Player's Chronicle, even citing the page number that the above diagram occurs on. In the text, he describes that the game ended in a draw, but that later joint analysis by the five players in the headline discovered a way for White to win. For completeness, here is the main line in Berger, which starts from the BCE position.

1.Ra7 1.f6+? is the subject of the BCE correction. This was apparently the game continuation, when 1...Kg6! 2.Ra3 leads to the first position above. 1...Bg8 2.Ra8 Bd5 3.Ra5 Bg8 4.Ra7 Bd5 5.Kd6+ Bf7 6.Ke5 Kg8 Fine chose 6...Kh6 as his main line 7.Kf6 Bd5 8.Ra3 Bb3 9.Ra8+ Kh7 10.Ke7 Bd5 11.Ra3 Bb3 12.Kf8 Kh6 13.f6! Kg6 14.f7 Bc4 15.Ra7! Kf6 16.Ke8! a1Q 17.f8Q+ 1-0 Despite working well in advance of tablebases, with the exception of Black's last desperate promotion (White mates in 17 after 16...Bxf7+ where 16...a1Q leads to mate in 9), the 19th century analysts managed to find both the fastest winning line for White and the move of longest resistance for Black.

Benko deleted the example in the revised edition, but kept Fine's text about the win being difficult against an advanced pawn.

6/20/20 - Bereolos-de Firmian, 2005 Kings Island Open

I've posted notes to my game against Nick de Firmian in the opening round of the 2005 Kings Island Open. I had given very light notes to it at the time, but have expanded on them now. It was a fairly one sided game, but the recapture decision on move 20 is instructive. My scoresheet indicates I only spent one minute on this move, so I was instinctiviely trying to improve my knight and not concerning myself with his. The difference between White's unstable Nd4 vs. Black's rock Ne5 gives Black a large advantage. In fairness, the game was played at a faster time control (G/75), but the fact that I hardly paused here makes me wonder if I would have played any differently with more time.

6/17/20 - BCE-600a, von Guretzy-Cornitz, 1864

Analysis of the ending queen vs. rook plus pawn dates all the way back to Cozio and Philidor in the 18th century. Kling and Horowitz added to the theory with some studies in their books in the middle of the 19th century. The next leap forward was by Bernhard von Guretzky-Cornitz in articles in Neue Berliner Schachzeitung in 1864. Fine used four of Guretzky-Cornitz's positions in BCE. One of these was the subject of a previous correction. Unlike that instance, Guretzky-Cornitz had the the assessment correct for this one, but slipped up in the analysis. Fine only included the main line, but Guretzky-Cornitz did consider the saving 5...Rc2

However, after 6.Qb5+ he only considered 6...Kc1? Black has to stay on the short side of the pawn with 6...Ka2 or 6...Ka1 (but not 6...Ka3? 7.Qb1! Rc3 and now White triangulates 8.Kd6 Rc4 9.Kc6 Rc3 10.Kd5 returning the move to Black and putting him in zugzwang. 10...Ka4 11.Qb2 forces the rook away from the pawns defense and 10...Rb3 11.Qc2 is no better.

Benko decided to remove this example for the revised edition, leaving only Fine's introductory text But if the White King can manage to attack the Pawn (as a result of a favorable initial position due to prior exchanges) the game is won. I thought this was an odd comment from Fine since the variation he presents begins with 1.Ke4 removing the attack on the pawn.

6/10/20 - BCE-388, Alekhine-Euwe, Berne 1932

Alekhine and Euwe played a number of rook endings of interest that appear in BCE. We've seen some from their World Championship matches previously, today's position is from the tournament in Berne 1932. This event was a combination international tournament and Swiss championship. The foreign players dominated taking the first six spots led by the World Champion, Alekhine, with a dominant 12.5/15 performance a full point clear of Euwe and Flohr. Euwe managed to hold their individual encounter despite being a pawn down in a rook ending after 36...Rxg5+

37.Kf3 The starting position of BCE-388 37...Ra5 37...Rf5+ is the subject of the BCE correction 38.a3 Fine shows that 38.d5 is also too slow. Black is able to grab enough pawns and get his own pawns rolling in sufficient time to hold the draw. 38...Rb5 39.Rd7 Kg7 40.Rxa7 Rxb2 41.Re7 Kf6 42.Re2 Rb3+ 43.Re3 Rb2 44.h4 Another try is 44.Re8 Fine's long variation appears correct leading to a draw as the rook on e8 is not very well placed. 44...Rd2 45.Ke4 Rxf2 46.Rb3 Re2+ 47.Kd5 Re6 48.Rc3 Ke7 49.Rc7+ Ke8 50.a4 g5 51.hxg5 Rg6! 52.Rb7 Rxg5+ 53.Kc6 Rg6+ 54.Kc7 f5! 55.d5 Rg7+ 56.Kxb6 Rxb7+ 57.Kxb7! Kd7! 58.a5 f4! 59.a6 f3! 60.a7 f2! 61.a8Q f1Q! 62.Qc8+ Ke7! King in front of the pawn is a general rule of thumb for the defender in queen + pawn versus queen, but here the king would be driven out of the defense after 62...Kd6? 63.Qe6+ Kc5 64.Qc6+ Kb4 65.d6 and White should win 63.Qe6+ Kd8! Here king in front is necessary, otherwise Black gets squeezed away after. 63...Kf8? 64.Kc7 Qc4+ 65.Qc6 and the pawn advances 64.Qd6+ Ke8 65.Qc6+ Kd8! 66.Qd6+ Ke8 67.Kc7 Qf7+! 68.Kb8 Qd7 69.Qe5+ Kf8 70.d6 Kf7 1/2-1/2

6/7/20 - TCEC Season 17 Superfinal, Games 11-12: French Winawer

I likely would have skipped over games 11-12 of the TCEC match since it featured a sideline of the Winawer variation that is not really on the cutting edge of theory 1.e4 e6 2.d4 d5 3.Nc3 Bb4 4.e5 b6

However, since Lc0 won game 12 in a 180 move marathon to even the score, it bears some attention. Stockfish again seemed to display some weaknesses in a closed position. After 19.Rh3

As in game 9, Stockfish chose to seal off the queenside with 19...a4 20.Nd3 c4 21.Nf4 b5 22.a3 Again, I'll speculate that Stockfish might be overweighing the fact that White has a "bad" bishop. Here, Black's bishop isn't much better. White eventually managed to trade the bishops after 42.N2xh4

White is undoubtedly better, but the question is how to make progress since Black has the f-file completely covered. Lc0 first widened the beachead on the kingside by bringing its king up and playing g5. 55.Kxg5

From here, Lc0 didn't demonstrate any winning idea, shuffling its pieces back and forth until 87.Rf2

Despite playing 32 moves, if we compare this to the last diagram, the White pieces could have achieved this configuration in as few as 9 moves. It seems like Black might be able to reach the drawing haven of the 50 move rule. However, Stockfish seemed to blow a fuse here with 87...b4?I don't have any good explanation for this move. Why shed a pawn for nothing when your opponent is just shuffling back and forth? I ran a version of Stockfish out for a long time and was surprised to see this move slowly bubbling its way to the top . I think the explanation is that Stockfish is evaluating that it is lost here, but like Lc0 doesn't have a convincing plan for White. So all of its choices are scoring about the same including b4 since after 88.cxb4 it is still hard to figure out how White wins. I don't know if it was absolutely necessary, but Stockfish allowed a trade of rooks 97...Qxe8

This began another series of shuffling culminating in 123.Ka2

Here, Stockfish again made Lc0's job easier by allowing a queen exchange with 123...Kd8?! 124.Qg5+ Qe7 125.Qxe7+ Kxe7 and after another 55 moves, Lc0 finally brought home the point. With the queens off, the White king can be much more mobile and he eventually penetrated via b4. Of course none of that would have been possible if Black hadn't needless given up his b-pawn. A very strange game!

6/5/20 - Rudolph-Bereolos, 1979 US Open

Another pawn ending from my early days of playing. I played the pawn ending correctly this time, but there were a few hiccups in the transition. This was from the US Open in Chicago against Alexey Rudolph, who is known these days as WIM Alexey Root. After 31.dxe4

Black has a rather useless extra doubled pawn. White's domination of the f-file gives full compensation. 31...Rc7 31...Qd4+!? is a computer move that should also be equal after 32.Kg2 Qc4 but it seems a bit reckless to leave f6 unguarded 32.Rxc7 Qxc7 33.Qf6 Qg7 34.Qxg5 The pawn ending after 34.Qxg7+ Kxg7 is drawn as there is no way for either side to penetrate. However, there is nothing wrong with regaining the pawn. It might even seem that White is a tad better because her queen is more active, but the White king is too exposed and there are numerous routes for Black to activate his queen: down the f-file as happens in the game, down the d-file, or via either the h6-c1 or a7-g1 diagonals. 34...c4 35.Qd8+ Kh7 36.Qd5 a6 37.Kg2 Qf6 38.Qd1 Black also makes some progress after 38.Qd2 Qh4 39.Qe2 Qg5 38...Qf4 39.Qf3? 39.Kh3! Qxe4 40.Qd7+ Kh6 41.Qd8 and it seems that the Black king cannot escape perpetual check

39...Qd2+? 39...Qxf3+ 40.Kxf3 Kh6! 41.Kg3 Kg5! wins as Black has more waiting moves on the queenside. Ironically, capturing the pawn on g5 on move 34 is what gave Black this opportunity as otherwise the Black king would not have a route into the White position. 40.Qf2? Qd6? again 40...Qxf2+ 41.Kxf2 Kh6! 41.g5 Resealing the kingside, now there shouldn't be too much left 41...Qe6 42.Kg3 Kg7 43.Qd2 Kf7 44.Qd8 Qe7 45.Qd2 Qc5 46.Qd7+ Qe7 47.Qd5+ Qe6 48.Kf3? 48.Qxe6+ is a simple draw

48...Qxd5! 49.exd5 Ke7! 50.Ke4 Black has so many spare tempi on the queenside it doesn't matter if White waits a move before going to e4 50.Ke3 Kd6 51.Ke4 b4 52.c3 (52.a3 c3-+) 52...a5 53.cxb4 axb4 54.a4 b3-+ 50...Kd6! 51.b3 The other pawn moves aren't any better 51.c3 a5 52.b3 cxb3 53.axb3 a4-+; 51.a3 a5 52.b4 (52.b3 b4 53.axb4 cxb3!) 52...axb4 53.axb4 c3 54.Kd3 Kxd5 55.Kxc3 Ke4 56.Kd2 Kd4-+ 51...cxb3! 52.axb3 b4 53.c4 bxc3! 54.Kd3 Kxd5 55.Kxc3 e4 56.Kd2 Kd4 57.Ke2 e3 58.Ke1 Kc3 59.Ke2 Kxb3 60.Kxe3 a5 61.Ke4 a4 62.Ke5 a3 63.Kf6 a2 64.Kxg6 a1Q 65.Kh7 Qa7+ 66.Kg8 Qg1 0-1

6/3/20 - BCE-321, Karstedt 1896

BCE-321 is a theoretical position that Fine references as being from Rabinovich. However, Rabinovich cites the analysis as coming from Berger and Karstedt, so that is likely where the error originated. The Encyclopedia of Chess Endings cites the position as a study by Karstedt from Deutsches Wochenschach in 1896, so that is what I am taking as the original source. Unfortunately, I was unable to find a scan of that year's issues to check for sure. ECE also indicates that the position is drawn with the White rook on b3 or b1. However, a separate ECE entry with the rook on b4, also attributed to Karstedt, is shown to be winning for White

1...Rc6 2.Kb5 Kc7 3.a7! Rb6+ 4.Ka5! Rxb4 5.a8Q! and unlike the positions with the rook on b3, b2, or b1, here Black does not have a check on the a-file to pick up the new queen.