Shakmaty Bereolos - The Official Chess Site of Peter Bereolos

Games versus GMs

Welcome to my chess page. This is mostly random thoughts and analysis in the form of a chess diary with other sections of the site slowly developing. A lot of the content will come from my own experience. There are two reasons for this. One, so I can use this site as a self-improvement tool. Two, so you the readers will have content that is not found on other chess sites. Follow the link to the left to reach my annotated games against grandmasters. Send me comments and ideas

If you link to this site, please point to so that your link will be correct no matter where I host the site.


Corrections to Basic Chess Endings

3/20/19 - BCE-399, Leonhardt-Schlechter, Nuremburg 1906

The 15th congress of the German Chess Association was a fairly strong tournament. According to Sonas it had the numbers 2-5 players in the world at the time (Tarrasch, Janowsky, Schlecter, and Marshall), but lacked Sonas' #1 Maroczy and the World Champion, Lasker, who Sonas only ranked as #6 at the time. The tournament didn't follow form at all. Marshall won 1.5 clear of Duras, Schlechter was tied for 3rd with Forgacs another half point back. However, Tarrasch (=9th) and Janowsky (tied for last) faired poorly in the 17 player event. The players might have been affected by the marathon tournament in Ostend, which finished shortly before this event where a multi-stage format was used with the top players playing 30 rounds! Tarrasch did not compete in Ostend, but the 3 others did and all made it to the final stage.

Ostend was a big success for Schlechter who won that tournament. He was pretty solid, as usual, in Nuremburg suffering only 1 loss, to tailender Janowsky. He could have also lost this week's BCE ending against Leonhardt, who was also near the bottom of the table. I thought this was a very instructive ending, but I didn't like Fine's presentation where he analyzes the position with both White and Black to move. Instead, he could have shown the mistakes by both players in the game by backing up one move to the start of the rook ending after 38.Rxa2

38...Kh7? Black should grab queenside space immediately with 38...a4! 39.g4 Kh7 40.g5 Kg6 41.h4 f6 with a draw 39.Kf2? Letting Black off the hook. This is position BCE-399 except that Fine places the rook on a1 instead of a2 (as does Benko), which makes no difference. Fine analyzes it with both sides to move, but the position of the king on f2 versus g1 does not seem to be relevant as Black does not have a good way to make use of the extra tempo. So White should win with similar play to that shown by Fine with 39.Ra4! Kxh6 40.Kf2 Kg5 41.Kf3 and Black will gradually be pushed background The game continued 39...a4! 40.Ra3 Kxh6 41.Kf3 Kg5 42.g3 Kf6 43.Kf4 Ra5 44.g4 Kg6 45.e5 Ra6 46.h4 1/2-1/2 Fine continues the analysis with 46...Kh6 46...f6 draws in similar fashion to the BCE correction 47.g5+ Kg6 47...Kh5 seems to immediately stop White's progress 48.Kg4 Ra5 49.h5+ Kg7 50.Kf5 Ra6! 51.h6+ Kh7 52.Rc3 a3 53.Rc7 a2 54.Rxf7+ 54.g6+ Kxh6 (54...Rxg6 55.Rxf7+ Kxh6! 56.Ra7! transposes) 55.Rxf7! Rxg6! 56.Ra7!= 54...Kg8! not 54...Kh8? 55.g6! Ra8 56.g7+! Kh7 (56...Kg8 57.Kg6 (or the fancy 57.Rf8+ Rxf8+ 58.Kg6!) ) 57.Rf8!+- 55.g6 Ra8! Fine stops here. White still has the try 56.h7+ Kh8 57.g7+! Kxh7 58.Rf8! Rxf8+! 59.gxf8Q! a1Q! and the Q+P vs. Q ending is a theoretical draw.

3/18/19 - Bereolos-Schneider, 2002 US Masters

One of the more complex rook endings I played was against Dimitri Scheider in the final round of the 2002 US Masters. I analyzed this ending extensively at the time, but there were many long variations and they often led to K+3P vs. K+R+P positions that hinged on a single tempo. Now that those positions are solved with the 7-piece tablebases, I've taken a fresh look at some of this analysis. Back then my conclusion was I'm not yet totally convinced that White is lost, but I haven't found a way to hold it yet. Now I think the conclusion is that the game never left the drawing zone.

A major focus of my analysis was on the variation that could have arisen after 47. h3

A critical position arises after 47...Kf4 In the game he retreated with 47...Rd5 48.h4 Ke3 48...Kxf3 49.g5= 49.h5 49.g5? loses as I showed back in 2002. 49...Kd4

White can go 3 directions with the rook 50.Ra3! 50.Rxd3+? loses fairly trivially as I originally analyzed 50...cxd3 51.h6 d2 52.h7 d1Q 53.h8Q Qa4#; while going backwards allows Black to keep his b-pawn and squeeze the White king 50.Rc1? Rb3+! 51.Ka5 Kc5! 52.Ka6 Ra3+ 53.Kb7 b4 54.h6 Re3 55.f4 b3 56.g5 b2 50...Rd1 I had previously dismissed 50...Rxa3!? 51.Kxa3! c3 52.h6! Kd3 53.h7 c2 54.h8Q c1Q+! with the conclusion and Black has only a slight edge in this queen ending, but it is quite dangerous for White 55.Kb4! Qc4+ 56.Ka5 b4

and now the computer says 57.Ka4! is the only drawing move. I doubt I would have found this, allowing the pawn to advance with check. (Instead, 57.Qxf6? b3! 58.g5 Kc2! 59.Qf5+ Kc1 60.g6 b2! 61.Qg5+ Kd1 62.Qg1+ Kc2 63.Qf2+ Kb3 64.Qe1 Ka2 65.Qf2 Ka1 and queens) 57...b3+ 58.Ka3! Qc3 59.Qh7+! and Black can't escape the checks 51.Kxb5! c3 52.Ra2! It is too soon to sacrifice 52.Rxc3? Kxc3 53.Kc5 Kd3 54.Kd5 Ke3+ 55.Ke6 Kxf3 56.Kxf6 Kxg4 57.h6 Rd6+ 58.Kg7 Kg5 59.h7 Rd7+ 60.Kg8 Kg6 61.h8N+ Kf6 and the knight is lost; 52...Rd2 53.Ra6 or to a7 or a8. This is the resource I did not find in my original analysis where I only considered 53.Ra4+? and 53.Ra1?

53...Rb2+ 53...Kd5 54.Kb6! This move looks strange cutting his own rook off along the rank, but the key is to clear the 5th rank to allow the White rook to stop the c-pawn while at the same time maintaining a path for the White king to cross to the kingside. (Instead, the king is in the way after 54. Kb4? c2! 55. Ra5+ Kd6! 56. Rc5 Rd4+! 57. Kb5 Rd5! when the rooks come off and Black queens) 54...c2 55.Ra5+! Kd4 56.Rc5!=; Black can also try immediately pushing the pawn 53...c2 54.Rc6! Kd3 55.Kb4 Rd1 56.Rc3+! Kd2 57.Kc4 c1Q 58.Rxc1 Rxc1+ 59.Kd4! shouldering the Black king (59.Kd5? Ke3 loses as above where White gets stuck with a knight in the corner) 59...Ke2 60.Ke4! Rc4+ 61.Kf5! Kxf3 62.h6!= 54.Kc6! other moves allow Black to trade rooks, but a key point now is that the Black king is denied access to d5. 54...c2 55.Ra4+! Kd3 56.Ra3+! Kd2 [56...Kc4 57.Ra5] 57.Ra1! Rb1 58.Ra2! Kd3 59.Rxc2! Kxc2! 60.Kd6

Black is a tempo too late to win now 60...Kd3 61.Ke6 Ke3 62.Kxf6 Kxf3 63.h6 Rb6+ 64.Kg7 Kxg4 65.h7 Kg5 66.h8Q Rb7+! 67.Kg8 Rb8+! with a draw. The trick from queen endings of 67...Kg6? doesn't work with a rook because the rook only has one checking square on the 8th rank, which White can cover. 68.Qe5

Lessons from this ending: 1. Rook endings are hard! 2. Sometimes you have to look at unconvential moves to achieve your aim. I'm thinking in particular of the line with 54.Kb6! 3. In queen endings, quality can be much more important the quality. 4. Also in queen endings, the king can be very useful as a defender against a passed pawn. (57. Ka4! and 58. Ka3!).

3/16/19 - Yilmaz-Saraci, 2018 Batumi Olympiad

Continuing the Olympiad/Yearbook 128 series with a game in the Vienna Variation, which corresponded to a survey by Nikolay Ninov. In the round 2 game between Mustafa Yimaz and Nderim Saraci on board 1 of the Turkey-Kosovo match, Black goes his own way in the opening, so the survey didn't prove to be particularly relevant.

1.d4 Nf6 2.c4 e6 3.Nf3 d5 4.Nc3 dxc4 5.e4 Bb4 6.Bg5 h6 7.Bxf6 Qxf6 8.Bxc4 c5 9.e5 Qd8 10.0-0 cxd4 11.Ne4

11...Nc6 11...0-0 12.Qe2 would transpose to the starting position of the survey where Ninov suggests that Black is fine after 12...Be7 Over the next few moves, Black could transpose to main lines by castling kingside, but in this game his idea appears to be to take the king in the other direction. 12.Qe2 Bd7 13.Rfd1 Qb6 14.a3 Be7 15.b4 0-0-0 16.Rac1 Kb8 17.Bd3 f6 18.Nc5 Bxc5 19.Rxc5 Ne7 This takes the pressure off of the White center allowing him to maneuver his knight towards d6. Black probably should have started with 19...Rc8 20.Nd2 Rc8 21.Nc4 Qd8 22.Nd6

Now White is clearly better and his attack proceeded very smoothly. 22...Rxc5 23.bxc5 Bc6 24.Be4 Qd7 25.Bxc6 Nxc6 26.Rb1 fxe5 27.Qa6 This looks even stronger than 27.Rxb7+ Now White's Queen will pick up the loose Black pawns after winning Black's Queen. 27...Nd8 28.Nxb7 Qxb7 28...Nxb7 29.c6 29.Rxb7+ Nxb7 30.Qxe6 1-0

3/13/19 - BCE-252, Schlechter-Walbrodt, Vienna 1898

This week starts a trio of BCE corrections featuring the games of Carl Schlechter. Schlechter was one of the top players in the early part of the 20th century. Sonas has him peaking at #2 in the world in late 1906 through early 1907, trailing only Maroczy, and ahead of World Champion Emanuel Lasker. He got his shot at the title in 1910, drawing Lasker 5-5, losing a titanic struggle in the final game. It has been much debated since that time if there was a clause that required Schlecter to win by two points in order to take the title.

Schlechter also drew matches with Marco (twice), Zinkl, Janowski, Alapin, Teichmann, and Tarrasch, which probably contibuted to his reputation as "the drawing master". My database shows 48.7% of his games ended in draws, which was quite high for that time. By comparison, Lasker's drawing percentage was less than half as much at 23.7%.

Vienna 1898 was a massive 20-player double round robin with all the top players in the world with the exception of Lasker and Charousek. Tarrasch emerged the winner in a tiebreak match over Pillsbury. Schlechter was a respectable 5th with a +7 score. His 17 draws were only 4th most in the tournament.

The subject of this week's post is Schlechter's knight versus bishop ending against Paul Walbrodt. A series of early exchanges landed them in the ending after only 23 moves 23...Kxf8

Despite the pawns on the queenside being fixed on the same color as the Black bishop, he should be able to hold this ending. 24.Nd1 Ke7 25.Ne3 Kf6 26.h4 Bf5 27.b4 Bd3 28.Kf2 h5 29.Kf3 g5 30.g3 Be4+ 31.Ke2 Kg6 32.Kd2 Kf6 33.Kc3 Bg6 34.b5 Be4 35.a4

35...gxh4? a grave mistake giving the White king access to the f4 square. I don't see a way for White to break through if Black just waits with 35...Bf5 when the bishop can retreat back to defend the queenside if necessary. 36.gxh4 Ke6 37.Kb4 Bd3 38.b6 a6 39.a5 White would love to preserve this spare tempo, but since the knight has to guard the f5 square until the White king can come around, the a4 pawn turns out to be a weakness after 39.Kc3 Bf5 40.Kd2? a5! 39...Be4 40.Kc3 Kf6 41.Kd2 Bf5 42.Ke1 Be4 43.Kf2 Ke6 44.Kg3 Kf7 45.Kf4 Kf6 46.Nf1 Bf5 47.Ng3 Bg4 48.Nh1 Be6 49.Nf2 Bf5 50.Nd1 Bg4 51.Nc3 Bd7 52.Na2 Be6 53.Nb4 Bc8 54.Nd3 Bf5 55.Ne5 Bc8 reaching BCE-252. The tournament book had already analyzed this position as a win for White, so it is odd that Fine did not just copy that analysis. 56.Ke3 Ke6 57.Ng6 It took Schlechter a while to hit upon the correct idea. 57.Nd3 would be faster 57...Kf6 58.Kf4 Bd7 59.Ne5 Bc8 returning again to BCE-252. Now White must be careful to avoid a triple repetition. 60.Nf3 Bf5 61.Ke3 61.Ne5? Bc8! is 3 times 61...Bc8 62.Ne5 Ke6 63.Kf3 Ke7 64.Nd3 finding the correct path to f4. It would be more efficient to start with 64.Ke3? to avoid the check on g4, but here 64...Ke6! would again be three times repetition. 64...Bf5 65.Nf4 Bg4+ 66.Ke3 Kf7 67.Nd3 Ke7 68.Ne5 Bc8 69.Kf3 Ke6 70.Nd3 Kf6 71.Ke3 1-0

3/11/19 - Fishbein-Bereolos, 1995 Kings Island Open

Although I am roughly the same age as GM Alex Fishbein, we never played as juniors and only seem to meet over the board one a decade. Probably a combination of geography and rating difference. I scored my first draw with Black against a GM against him at the Kings Island Open in 1995. It looks like I could have gotten more, but was unable to finish the job.

3/6/19 - BCE-222a

This week completes the triplet of positions from Kling and Horwitz with a more practical material distribution of N+P vs. B. Here, Black has a stalemate resource first discovered by Cheron in 1952. Fine compounded the error in the position by calling the incorrect first move by Black forced.

I thought I recognized the position as an excercise from Dvoretsky. I was mostly correct, except that he shows the mirror image position on the queenside, while still attributing it to Kling and Horwitz. This was shades of the curious position from BCE-226b. I didn't do a whole lot of research trying to track this down as the first source I checked appeared to be where the reversal occured. In the second volume of Averbakh's endgame series he gives the mirror image position, but cites it as from Kling and Horwitz with flanks reversed in a footnote. Since it is basically the same study, I don't think flipping the position even dropping the note as Dvoretsky did is too troublesome as long as the original composer is still getting the credit. The somewhat bigger mystery is why Averbakh reversed the flanks on this and other positions. It seems that he may have been trying to systematically look at the pawns on each file without having to jump back and forth from kingside to queenside in his examples. In the N+P vs. B section almost every example has a White pawn on the queenside or a black pawn on the kingside. The lone exception is the very last example in the section. The section on B+P vs. N has a similar configuration although in that section there are even more exceptions. I also don't see any examples in the B+P vs. N section where he called out having reversed the flanks.

3/5/19 - Lekau-Sharikhan, 2018 Batumi Olympiad

There are a couple of dangers in picking games to annotate based on an opening variation. The first is that the set of games you are choosing from does not contain that variation. We've seen this a couple of times in my Olympiad/Yearbook 128 series. When this has occurred, I think I've been able to select interesting games that are close to the variation in question. The bigger danger is that there is a game with the variation of interest, but it is not a very good game. That is the problem I had when looking for games to match Krisztian Szabo's survey on the Steinitz Poisoned Pawn variation in the French Defense. The only game with this line was played between two lower rated players, Romokotjo Lekau of Lesotho and Shawal Sharikhan of Brunei. I debated whether to try to choose a different game, but in the end I decided to go with this one and have an easy annotation day.

1.e4 e6 2.d4 d5 3.Nc3 Nf6 4.e5 Nfd7 5.f4 c5 6.Nf3 Nc6 7.Be3 Qb6 8.Qd2 cxd4 9.Nxd4 Qxb2

10.Nb3?? Perhaps White got confused with the Sicilian Najdorf Poisoned Pawn 1. e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.Nc3 a6 6.Bg5 e6 7.f4 Qb6 8.Qd2 Qxb2 where 9.Nb3 is a legitimate alternative to 9.Rb1. Szabo's survey focused on 10.Rb1 Qa3 11.Bb5 Nxd4 12.Bxd4 a6 13.Bxd7+ Bxd7 14.Rb3 Qe7 15.Rxb7 Rc8 (15...Qh4+ and 15...Qd8 are important alternatives) 10...Bb4 and just like that the game is over before it has really begun. 11.Qd1 Qxc3+ 12.Kf2 Nc5 13.Bb5 Nxb3 14.cxb3 Bc5 15.Re1 Bxe3+ 16.Rxe3 Qc5 17.a4 Bd7 18.Rc1 Qb6 19.Qg4 g6 20.Qh4 Ne7 21.g4 Bxb5 22.axb5 d4 23.Rd3 Nd5 24.Kg3 Qxb5 25.Rc4 Qa5 26.Rc1 Qb6 27.Rcd1 Rc8 28.Kf3 Rc3 29.f5 after this move the engines start announcing mate. 29...Qxb3 30.Ke4 Rxd3 0-1

older material

Stats since June 1, 2006