Shakmaty Bereolos - The Official Chess Site of Peter Bereolos

9/30/19-Wojtkiewicz-Bereolos, 1996 Kings Island Open

My first attempt against Aleks Wojtkiewicz's fianchetto variation against the Kings Indian was 10 years before the debacle at the Chicago Open. This one did not make it into Wojo's Weapons Volume 2, but very well could have. The first chapter is on the classical variation, which the authors call The Zurich 1953 Defense since the variation was played often in that famous Candidates Tournament. The first subsection is Black's Pawn on d6 Falls. That is what happened in this game and after that it was just a mop up.

9/28/19 - Matlin-Bereolos, Hammond 1983

Since I've been looking at N vs. P endings as part of the recent BCE posts, I had a look through my own history of these endings for interesting or instructive examples. Here is one of the earlier ones I found, versus Harry Matlin at a one-day tournament at the Hammond Chess Club. It looks like they named their tournaments after World Champions that year. I played tournaments named Mikhail Tal Saturday Sacrifices, Boris Spassky Fall Fishery, and this one was Anatoly Karpov Winter Winners. After 31.Rxd6

I swallowed the poisoned g-pawn 34...Nxg2? 35.Rxe6 White can't immediately take the knight 35.Rxg2?? Rxg2 36.Rxe6 Rg1+ 37.Kh2 R7g2# 35...Rxe6 36.Bc4 taking the knight is still bad 36.Rxg2? Rxg2 37.Bc4 (37.Kxg2 Rxe2+) 37...Rgg6 Now the pin costs Black a piece 36...Rgg6 37.Rxg2 Kh7 38.Rxg6 Rxg6 39.Bd3 c4 40.Bxg6+ Kxg6

Black could safely resign here. He can't even make a passed pawn because of the doubled b-pawns. White should just bring his king into the game to stop any penetration by the Black king. Instead, he goes pawn hunting with his knight. 41.Nd2 41.Kg2 Kf5 42.Kf2 Ke5 43.Ke2 Kd4 44.Kd2 and Black will have to give ground as White has infinite waiting moves with Nf1-h2-f1 41...b5 42.Ne4 Kf5 43.Nd6+ Ke6 44.Nxb5 Kd5 45.Kg2 Kc5 46.Nc3 Kd4 47.Kf2 Kd3 48.Nd5 Kc2 49.Nxf4 Kxb2

White is still winning, but Black has a small bit of hope. 50.a4 Kb3 51.Ke3 Kxa4 52.Nd5 b5 53.Kd4 53.f4 b4 54.f5 c3 55.f6 c2 56.Kd2 Kb3 (56...b3 57.f7) 57.Nxb4 53...Kb3 54.f4 b4 55.f5 c3 What should have been a trivial win has now become an calculation exercise.

56.Ne3? taking aim at the wrong square. He should be targetting c1 instead of c2 making Black spend an extra tempo with his king 56.Nf4! Kb2 (56...Ka2 57.Nd3 b3 58.Kxc3) 57.f6 c2 58.Nd3+ Kb1 59.f7 b3 60.f8Q b2 61.Qb4 +- 56...c2! 57.Nxc2! Kxc2! 58.f6 b3! 59.f7 b2! 60.f8Q b1Q! 61.Qf5+ Black makes perpetual check rather easily after 61.Qxh6 because of the offsides postion of the White queen, but that was the only remaining chance. 61...Kb2 62.Qxb1+ Kxb1! Perhaps White had evaluated the pawn ending as winning, but with a rook pawn, Black doesn't need to get his king in front of the pawn if he can trap the White king in front of it 63.Ke4 Kc2 64.h4 h5 65.Kf4 Kd3! 66.Kf5 Ke3 67.Kg5 Ke4! 68.Kxh5 Kf5! 69.Kh6 Kf6! 70.Kh7 Kf7 71.h5 Kf6 72.h6 Kf7! 73.Kh8 Kg6 74.h7 Kf7! 1/2-1/2

The main lesson from this ending is that the simplest way to victory can often be just eliminating any possible counterplay from your opponent.

9/25/19 - BCE-119b, Kieseritzky, 1842

Completing the trio of knight endings is a study by one of the top players of the 19th century, Lionel Kieseritzky. While data is sparce for that era, Sonas ranks him #1 for a 2 year stretch in the middle of the century. Despite his successes, he is probably best known for losing The Immortal Game to Adolf Anderssen.

Today's study first appeared in one of the first, if not the first, chess magazines, Le Palamède, revue mensuelle des √©checs et autres jeux (Monthly chess and other games). Notation has come a long way since that time. Today, we write the key move of the study as 1. Kd3!. Back then, it was 1. Le R à la 3 c. de la D. Nevertheless, the analysis was almost spot on. The only flaw I found was in the variation 1.Ne4? Ke6! 2.Nxc5+ Ke5 3.Nd7+ Ke6! 4.Nf8+

Here, Black must continue attacking the knight with 4...Kf7! Instead Keiseritzky's 4...Kf6? gives White the tempo he needs to activate his king with 5.Kd3 Kf7(5...b5 6.Ke4) 6.Nd7 Ke6 7.Nc5+ Ke5 8.Ke3 b5 9.Ne4 b4 10.Nxg5 b3 11.Ne4 b2 12.Nc3

The improvement 4...Kf7 had already been published in Berger, which was Fine's primary source. After 1. Ne4? the move 1...c5? did not appear in either Le Palamède, nor in Berger, so Fine may have been the original source for this error. The Encyclopedia of Chess Endings does not mention 1...c5 either, but does give 1...Ke6! an exclamation point. I did not research it further.

9/24/19 - Korbov-Le, 2019 World Cup

Since we have been on the subject of knight endings for the weekly BCE posts, I shouldn't let the knightmare endgame that arose in the World Cup game between Anton Korobov and Le Quang Liem pass unnoticed. In the double knight ending, Black was approaching victory, when the following occurred after 59. d7

The players were down to a few minutes each plus the increment and Le played 59...Kc6? 59...Ne6+! 60.Kf7 Nd8+! 61.Ke8 Nb7 62.d8Q (62.Na5 Ned6+! 63.Kf8 (63.Ke7 Nxa5) 63...a3) 62...Nxd8 63.Kxd8 Kd4 and the pawns should roll 60.d8N+ 60. d8Q? Ne6+ was Black's point. Now we have an endgame with 3 knights versus 2 which may be unique in the history of tournament chess. I found no other examples in the database 60...Kc5 61.Nb7+ Kb4 62.Nba5

This strange picket line of knights seems to hold the draw. 62...Kc5 63.Kf7 Nd5 64.Nxd5 Kxd5 65.Nb6+! Kc5 66.Nxa4+ Kb5 67.Nb2 Kxa5 68.Nc4+ Kb4 69.Nxe5 fxe5 70.Ke6! Nd6 71.Kxe5! 1/2-1/2

As I said, I was unable to find any further examples. However, I did find a few games where one player had only 3 knights plus pawns. The one I found the most strange was the high level game between Gabriel Sargissian and Levan Pantsulaia at the at the 2004 Aeroflot Open. After 44...Ng7

According to the game score in ChessBase and elsewhere, Sargissian opted for the underpromotion 45.f8N!? Threatening Nf7#, but 45.f8Q is more forcing and better. Now, the game score concludes 45...Ne8??either 45...Nxe6 or 45...Nd6 looks much better for Black. 46.Ne4?? I am sure that if a player of Sargissian's strength underpromoted on move 45, he would not have missed 46.Nf7#! 46...d3? 47. Ke3? 47.g5+ Kh5 48. Nf6+ Kh4 49. Nxg6+ picking up the rook is simple. 47...g5 1/2-1/2

So what really happened here? I suspect that Black actually played 45...Rxf8+ However, after White recaptures 46.Nxf8, if we continue along the moves of the database, then after47...d3 47. Ke3 g5 it seems like White would still have good chances to push for a win with his extra pawn after 46.Kxd3. I guess this will remain a chess mystery for now.

I wanted to share one fantastic variation I found trying to puzzle this all out. After 45...Rxf8 46.Nxf8If Black defends the mate threat with 46...Nd8 and we follow the database moves with 47.Ne4 d3 48.Ke3 g5 White would have the incredible shot 49.Nd6!!

Black is completely paralyzed. If Nd8 moves then White has Nf7#, while if Ng7 moves, White has Nf5#. Black can only make a few pawn moves until the zugzwang is complete. 49...a5 50.Kxd3 a4 51.bxa4 and mate next move.

9/18/19 - BCE-105a, Fine 1941

This week's BCE correction appears to be an original composition of Fine's. Generally, a knight can hold against two connected passed pawns unless the defending king is not participating. The correction is a good illustration of constructing a barrier as the improvement keeps the black king from penetrating.

Even when the king is far away, the agility of the knight can cause technical difficulties. At normal time controls, I would not expect the great players in the following example to misplay such an ending, but in blitz anything can happen. In Ivanchuk-Leko from the 2008 Tal Memorial Blitz tournament after 46. b5

Leko decided to try to hold the knight versus pawns ending 46...Qxc3? 47.Qxc3! Nxc3 48.b6! Distracting the Black king. 48...Kd7 49.h4 Ne4 50.g4 Kc6 51.g5? It is well known that knights have the most problems with rook pawns. So White should push the h-pawn first. 51.h5! Kxb6 52.h6! Ng5 53.Kg3! Kc6 54.Kh4! Nf3+ 55.Kh5 Ne5 56.h7 Nf7 57.g5 and the pawns roll through51...Kxb6! 52.g6 Nf6 53.Kg3 Kc6 54.Kf4 Kd6 The Black king has gotten back into play and Black holds the draw 55.Kg5 Ke7 56.h5 Ne4+ 57.Kf5 Ng3+ 58.Kg4 Ne4 59.Kf5 Ng3+ 60.Kg5 Ne4+ 61.Kh6 Kf8 62.g7+ Kf7 63.Kh7 Nf6+! 64.Kh8 Ng8! 1/2-1/2

9/15/19-Wojtkiewicz-Bereolos, 2006 Chicago Open

I've added my game versus Aleks Wojtkiewicz from the 2006 Chicago Open to the GM games section. I played a dubious opening variation in this one and while I fought for a long time, I never managed to overcome this bad start.

9/11/19 - BCE-111a, Troitsky 1906

I don't have many knight endings in the BCE corrections section, so I thought I'd remedy that. This week's example is fairly esoteric, with 2 knight battling 3 connected passed pawns. The winning idea for the side with the knights is to blockade the pawns, capture 2 of them, then win the ending of 2 knights versus a pawn. While this can theoretically be done, in practice it is difficult and can often run up against the 50-move rule.

Today's BCE position is drawn only because of the 50-move rule, which may cause some to question if it is really a correction. However, if we ignore the 50-move rule, then on the 2nd move both sides go wrong from an objective evaluation, so the position does merit inclusion. We saw this conflict between the tablebase evaluation and the 50-move rule in the game Fawzy-Chaparinov which was the very first post of my Olympiad/Yearbook 128 series. Things are a bit less clear in this ending because of the multiple pawns which reset the 50 move count every time one moves or is captured. However, I think I demonstrate in the correction link that it applies here too.

Of course, it doesn't take long for the name Troitsky to appear when the topic is 2 knights versus pawns. Fine doesn't give a citation for this position, but Troitsky analyzed it in an article in the June 1906 edition of Deutsch Schachzeitung. Fine follows Troitsky's analysis for quite a few moves before going his own way.

Of course Troitsky was not worried about the constraints of the 50 move rule, but he predated Fine on getting the objective analysis wrong on move two. 1.Kf4 e5+ Troitsky also looks at 1...c5 2.Ne4+ Ke7! 3.Kg5 which Black would hold with the 50-move rule and 1...d5? 2.Ne5! Ke7 when 3.Nxc6+ is a win for White, instead Troitsky continued (3.Na4? when Kd6! is again drawn with the 50-move rule) ] 2.Kg4?! Ke6?! Troitsky also considers 2...d5! 3.Ng5 e4 4.Kf4 Ke7?! (4...e3! is the only move that is objectively drawing, but the text is still a draw in 50-move rule space) 5.Ke5 e3? Strangely enough, this push, which was the saving move on the previous move when it could be captured is a losing move here. 6.Ne6 e2 7.Nxe2 c5 8.Kxd5! (Troitsky also indicates 8.Nxc5?! which allows 8...d4! 9.Nd3 (!) which is mate in 58, so the 50-move rule would again come into play) 8...c4 and although the pawn has crossed the Troitsky line, White has mate in 49. One must remember that the Troitsky line indicates the line for any king position. There are many cases where the pawn is over the line, but the defender is still lost as is the case here beginning with 9.Nc3!] 3.Ng5+ Kf6! 4.Nce4+ Ke7! 5.Kf5 Kd7? Fine's 5...d5! is the only move to hold (using the 50-move rule) It is a bit odd that Troitsky did not consider that move especially since he looked at it on earlier moves. The only alternative he gives here is 5...c5? 6.Nf6 c4 7.Nge4 +-6.Nf7 d5 7.Nc5+ Black loses his e-pawn and will be left with a pawn on c6 or d5 in the subsequent 2 knight versus pawn ending, both of which are behind the Troitsky line and the second Troitsky line, so White wins.

This ending is very rare in practice. I only found a dozen examples of 2 knights versus 3 connected passed pawns in the database. In two of those, the pawns were far advanced and beat the knights. In another, one of the knights was immediately captured. Of the other 9, the side with the knights won twice, once when the defending king was already in a mating net when the ending started and once using the technique described earlier of blockading the pawns and eventually winning the 2N vs. P ending. Even in that game the defender had multiple chances to draw both objectively and by reaching the 50 move rule. Since I ran through variations like that in the BCE link, I'll skip repeating that kind of tedious analysis here.

9/7/19 - Youssef-Bereolos, 1983 Region V Junior Invitational

Back in my junior days, the USCF was divided into regions. I don't really find much evidence that these regions still exist for the modern US Chess. Region V consisted of Indiana, Michigan, Ohio, and Kentucky. Every year the top 2 juniors from each of these states was invited to a Scheveningen style round robin where each player would face all the players outside of his state. The 1983 edition in Ann Arbor, Michigan was a big success for Indiana as Billy Colias and I each defeated future GM Ben Finegold and all of our other opponents for a shared first with 6-0 scores.

In the second round, I had Black against Michigan's Issa Youssef. Here is another pawn ending where the engines reveal flaws. I had found the saving draw for White, but my analysis did not uncover the fact that the ending was a win for Black. After 33...Kg8

White has an extra exchange, but is lost because of the passed e-pawn and ideas of ...Nf3+ or ...Nd3. 34.Qf4 probably the best practical chance, 34.Qd8+ (34.Qe6+ Kg7 35.Qe7+ Kh6 36.Qf8+ Kh5 just activates the Black king) 34...Kg7 35.Qxa5 leaves the White king too exposed 35...Qe3+ 36.Kg2 (36.Kh1 Nd3 37.Qc3+ Kh6 and White can't do anything about the threat of ...Qf3+ followed by ...Qf2+) 36...Nd3 37.Qc3+ Kh6 38.Kh3 trying to sidestep ...Qf2+ 38...Qe6+ 39.Kg2 Qe4+ 40.Kg1 (40.Kh3 Nf2#) 40...Qe3+ 41.Kg2 Qf2+ 34...Nf3+ 35.Kf2 Qxf4! 36.gxf4 Nxe1! 37.Kxe1

37...Kf7? Black needs to fix the White queenside pawn structure giving himself two spare tempi in the form of b6 and b5 37...a4! 38.Kxe2 Kf7 39.Kd3 Ke6 40.Ke4 (Running to the queenside is too slow 40.Kc4 Kf5 41.Kb4 Kxf4 42.Kxa4 g5 43.b4 g4 44.Kb3 h5 45.a4 h4 46.a5 g3 47.hxg3+ hxg3 48.b5 g2 49.a6 bxa6 50.bxa6 g1Q) 40...Kf6 41.Ke3 Kf5 42.Kf3 h6 43.h3 (43.Kg3 Ke4 44.Kg4 Ke3 45.Kg3 h5 and White gets outflanked) 43...g5 44.fxg5 Kxg5 45.Kg3 b6! 46.Kf3 Kf5 47.Ke3 Ke5 White has to pick a direction, but he loses the race no matter which side he goes to 48.Kf3 (48.Kd3 Kf4) 48...Kd4 49.Kg4 (White has only lost further ground after 49.Ke2 Ke4) 49...Kc4 50.Kh5 Kb3! 51.Kxh6 Kxb2! 52.h4 Kxa3! 53.h5 Kb3 the point. Black will queen first and control h8. 38.Kxe2? White needs to stop ...a4 with 38.b3 or 38.a4 38...Ke6? 39.Ke3? Kd5? missing the last chance for 39...a4! 40.b3! a4 41.bxa4! Kc4 42.a5! 42.Ke4? Kb3! and Black wins in similar fashion to the game lines. 42...Kb5 42...Kb3 43.Kd4! Kxa3 44.Kc5! Ka4 45.Kb6! Kb4 46.Kxb7 Kxa5! is a worse version of the next note from Black's perspective, but still a draw.

43.Ke4? White miscalculates the ensuing race. Instead, he should have drawn by counterattacking the queenside 43.Kd4! Kxa5 44.Kc5! Ka4 45.Kb6! Kxa3 46.Kxb7! Kb4 47.Kc6 Kc4 48.Kd6 Kd4 49.Ke6 Ke4 50.Kf6! Kxf4 51.Kg7! h5 52.Kxg6! h4 53.Kh5! h3 54.Kh4! 43...Kxa5! 44.f5 It is too late for 44.Kd5 as White has to spend an extra move to capture the b-pawn. 44...Ka4! 45.Kc5 Kxa3! 46.Kb6 Kb4 47.Kxb7 Kc5 48.Kc7 Kd5 49.Kd7 Ke4! 50.Ke6 Kxf4! 51.Kf6 g5 52.Kg7 h5! 53.Kg6 h4 54.Kh5 h3 55.Kg6 g4 56.Kf6 g3 44...gxf5+! 45.Kxf5 Ka4 46.Kg5 Kxa3 47.Kh6 b5! 48.Kxh7 b4! 49.h4 b3! 50.h5 b2! 51.h6 b1Q+! 52.Kh8 Qb8+ 53.Kg7 Qc7+ 54.Kg8 Qd8+ 55.Kh7 Qf6 0-1

Lessons from this ending: 1. Be alert for chances to gain extra tempi in a pawn ending (37...a4!). 2. Capturing is not mandatory (38. Kxe2?) 3. All variations in a pawn ending must be calculated (43. Kd4!).

9/4/19 - BCE-70, Lasker and Reichhelm, 1901

Lasker also made an impression in the world of chess studies. This week's BCE position is a classic example of corresponding squares. There seem to be multiple versions of this study. The van der Heijden database gives 3 variants. That was the starting point of my research into the postion, where I found some discrepancies. The original appears to come from Lasker's column in the Manchester Evening News where the April 10, 1901 edition gives the following position Composed by E. Lasker

Later that same year, Lasker toured the US. This seems to be the point where Gustav Reichhelm enters the picture. The position in BCE-70 is widely referenced as Lasker and Reichhelm from the Chicago Tribune. The Encylopedia of Chess Endings even attributes it entirely to Reichhelm. However, in the May 26, 1901 edition of the Tribune, it is the following position that appears.

This is the BCE position except the f-pawns are on f5 and f6 instead of f4 and f5. The text accompanying the position states During the champion's half hour visit at the Chicago Chess and Checker club last Tuesday he showed the following ending, his own composition, with a little alteration by G. Reichhelm to increase the difficulty, consisting of placing white's K at [a1] instead of [a3] and black's K at [a7] instead of [a8].

So where does the BCE position come from? It appears with the title Lasker's Great End-Game in the June 22, 1901 edition of The Literary Digest. As the name implies that is a compendium of information from different sources. However, they don't cite exactly where they pulled the study from, so maybe there is yet another source with this variant.

Finally, the HvdHDB gives one more version attributed only to The Literary Digest 1901 with no date given. Here, the kings are back in the position of the original study, but the f-pawns have shifted. However, I was unable to locate this position in The Literary Digest or elsewhere.

Amazingly, none of these subtle changes of king and pawn positions ends up impacting the assessment. They are all White to play and win.