07 August 2011

Annotated Game #4: GM Alex Yermolinsky simul

The following game was played at the 2002 National Open in Las Vegas.  This was my first simultaneous exhibition game and I was very much interested in playing against GM Yermolinsky, due both to his record in the U.S. Championship (having won twice in the 1990s) and his authorship of the excellent book The Road to Chess Improvement (Gambit Publications, 1999) which I had recently read.

Similar to what occurred in my game vs. GM Walter Browne, I was able to successfully prepare the opening prior to the game.  I determined that Yermolinsky was likely to play 1. d4 and that we would therefore end up in a Slav Defense.  The Geller Gambit (5. e4 in the main line Slav progression) seemed to suit Yermolinsky's style, especially in a simul, and he had played it before.  This swashbuckling gambit line is White's primary choice for aggressive play, at least in the main lines of the Slav, and Black has to know it cold.  However, if Black rides it out, he gets a middlegame position which is at least equal and contains no real dangers.

My primary source for opening preparation was Graham Burgess' entertaining and thorough book The Slav, published by Gambit in 2001.  This has been an essential opening reference for me ever since, although it demands effort from the reader, since it focuses more on being a comprehensive theoretical treatment of the opening rather than on demonstrating basic concepts.  However, its depth and the author's ability to convey in words his assessment of various lines and approaches make it accessible for my opening study methods.

When preparing the opening, although I certainly made the effort to memorize the moves themselves, what really helped my ability to then play them over-the-board was reviewing and memorizing the concepts behind them.  The Geller Gambit for Black can be broken down into a series of mini-operations from move 8 onward:
  • After 8 axb4, exchange off the White knight on c3 first in order to be able to recapture with the b-pawn
  • Develop the light-squared bishop to b7 after Ng5, seizing the long diagonal
  • g6 is the obvious (and only) defensive move after Qh5
  • After the White queen retreats, hit the knight on g5 with Be7, then bring out Nd7
  • When White opposes his bishop on the long diagonal, move Qc8 to protect b7 rather than exchanging
  • Castle and then play f5 to exchange down material
That takes Black through move 18 and gives him a position that is at least even and likely contains any winning prospects that may exist.  When playing this opening, White essentially gambles that Black will go astray sometime before move 16.  Although I make some inferior moves in the game after the opening, it is completely drawn and without any prospects for White until move 36, then it quickly becomes a winning game for White after move 38.  What happened?  Following the erroneous beginner concept of simplification = good, I gave up first the double rooks and then the single rooks, leading to a winning K+P endgame for White.  GM Yermolinsky pointed out to me afterwards that keeping both rooks on would have completely stymied White, who had no way to make progress.

With this game, I was pleased that the opening preparation had worked so well, although I was disappointed with the final result, which was due to my lack of endgame knowledge.  That said, the valuable lesson about the effects of endgame simplification is something that will stay with me.

[Event "Simul"] [Site "?"] [Date "????.??.??"] [Round "?"] [White "Yermolinsky, Alex"] [Black "ChessAdmin"] [Result "1-0"] [ECO "D15"] [Annotator "ChessAdmin/Fritz/Houdini"] [PlyCount "93"] [EventDate "2002.??.??"] {D15: Slav Defence: 4 Nc3 a6 and gambit lines after 4 Nc3 dxc4} 1. d4 d5 2. c4 c6 3. Nf3 Nf6 4. Nc3 dxc4 5. e4 {the start of the Geller Gambit} b5 {anything else simply gives White a free hand in the center with no compensation} 6. e5 Nd5 7. a4 e6 8. axb5 {this is the first real decision point for White, with Ng5 being a possible alternative.} Nxc3 9. bxc3 cxb5 10. Ng5 Bb7 11. Qh5 g6 12. Qg4 {this or Qh3 is necessary in order to protect the g2 pawn and free the light-square bishop for development.} Be7 13. Be2 Nd7 {Bd5 is an alternative, this is considered the main line.} 14. Bf3 Qc8 ({At this point, trading material would make it harder to defend, as} 14... Bxf3 15. Qxf3 O-O 16. h4 { gives White the h1-a8 diagonal and good queen placement.}) 15. Ne4 O-O { Essentially the end of the book lines. Now Burgess gives h4 for White, while my database shows only one game, which follows the actual continuation.} 16. O-O f5 {this is the essential defensive move in alternative lines for White and here as well, forcing liquidation of material.} 17. exf6 Nxf6 18. Nxf6+ Rxf6 19. Be4 {here Yermolinsky stops following Hodakowsky-Hinne (1959 West German championship), the game from my database, where White exchanged on b7.} ({If} 19. Bg5 Bxf3 {followed by Rf7 is the best response, White now does not have enough material to press an attack.}) 19... Bd5 {Fritz recommended Bxe4 here; Houdini also likes this but considers a5 superior, with some advantage to Black as he mobilizes his extra queenside pawn. The text move, although it loses material, results in a dead even position.} 20. Bxd5 $11 exd5 21. Qxc8+ Rxc8 22. Rxa7 Re6 (22... Rf7 {preparing to take away the 7th rank looks better. This is where my lack of endgame ability starts to show. Sample continuation by Houdini is} 23. Rb7 b4 24. cxb4 c3 {where Black's advanced pawn will tie White up.}) 23. Rb7 Re2 {here it would have been better to implement the idea from the previous varation, b4 followed by c3, but I did not see this concept at the time.} 24. Be3 Ba3 25. Rxb5 Rc2 26. Rxd5 Rxc3 27. Rd7 Rd3 {Although Black is a pawn down, the double R+B ending is still drawn, thanks to the c-pawn and activity of Black's pieces.} 28. Bc1 ({Better is something like} 28. Bh6 Bb2 29. Rg7+ Kh8 $11) 28... Bxc1 ({Both Fritz and Houdini saw the improvement} 28... Bc5 $15) 29. Rxc1 $11 c3 30. Kf1 Rd2 31. Re7 Rxd4 (31... c2 32. Ke1 Rxd4 33. Re2 $15 {was Fritz's evaluation, but Houdini considers it drawn and I didn't consider it possible to get anywhere by pushing the pawn, either.}) 32. Re3 Rdc4 33. Rc2 Kf7 34. Ke2 Kf6 35. Kd3 Kf5 {Although not losing, moving the king away from protecting the kingside pawns and the 7th rank is not optimal. Houdini suggests simply retreating the c4 rook, which will not allow White to make any progress.} 36. Re8 h5 {This is a key turning point. Keeping it a double-rook endgame would make it impossible for White to make any progress.} ({Both Fritz and Houdini prefer} 36... R8c6) 37. Rxc8 $16 Rxc8 38. Rxc3 {At this point the rook endgame should still be drawn. However... } Rxc3+ {and it's now a winning K+P endgame for White.} (38... Rd8+ 39. Ke3 g5 {should hold}) 39. Kxc3 $18 Ke4 40. h4 Ke5 41. Kd3 Kf5 42. Kd4 Kf4 43. Kd5 Kg4 44. Ke5 Kxh4 45. Kf4 g5+ 46. Kf5 g4 47. Kf4 (47. Kf4 g3 48. fxg3#) 1-0


  1. Wow, that was some very good preparation. I played this variation several times back in the day and luckily no one was this ready! I seem to recall a game between Smyslov and Petrosian where white almost blew away black, but I can't recall who was playing which color...

  2. I looked it up, it's Petrosian-Smyslov, 1951 Soviet Championship. Petrosian varied from the posted game with 14. h4, after which play continued ..h5 (the only move) 15. Qg3 Nb6 16. O-O a5. This last pawn push looks a bit suspect to me, especially since Smyslov decided to castle queenside on move 19. White shortly afterwards got his knight to d6 and Black lost the exchange removing it (and the game along with it).


Your comments and ideas on chess training and this site are welcomed.

Please note that moderation is turned on as an anti-spam measure; your comment will be published as soon as possible, if it is not spam.