Monthly Archives: October 2015

When to take a draw ?

Often I lose a game after rejecting an offer of a draw. I think this is mainly because I tend to overreach in good positions when short of time, rather than because I overestimate my position or am wrongfooted by the offer. On the whole I think it’s good practice to play on if one thinks one has the better position or there are interesting chances for both sides. But there are times when I wonder.

My season started with three tests of my recently-adopted Benko Gambit (I am playing Black in each case). Repetitions or draw offers played a part in each of them.

In the first, played for Pimlico, I get a promising position, refuse the draw offer, and suffer for my principles:

You need to activate javascript to enhance chess game and diagram visualization.
1. d4 Nf6 2. c4 c5 3. d5 b5 4. cxb5 a6 5. f3 {The Dlugy. For once, I had done some preparation and expected this.} 5... axb5 6. e4 Qa5+ 7. Bd2 (7. b4 {is an interesting gambit.}) 7... b4 8. Na3 d6 9. Nc4 Qd8 10. Bd3 e6 ({One of the spectators asked why not} 10... g6 {which is reasonable, but White&#8217;s control of c4 makes this approach less attractive for Black and the books recommend immediately challenging the centre with e6.}) 11. Ne3 exd5 12. exd5 Bb7 {White&#8217;s control of c4 is roughly balanced by the chronic weakness of d5.} 13. Qb3 Nbd7 14. Ne2 Be7 15. O-O Qc8 16. a3 (16. Rfc1 {and the threat of Bxb4 makes my last move look rather silly.}) 16... bxa3 17. Rxa3 Rb8 18. Qa2 O-O 19. Nc3 Ne5 20. Bf5 {I hadn&#8217;t considered this} 20... Qc7 21. Ne4 (21. f4 {may give an edge}) 21... Ra8 {Making space for the queen.} 22. Ba5 Qb8 23. Bc3 Rxa3 24. bxa3 Nxe4 {Nine minutes here looking for something forcing.} 25. fxe4 Bf6 26. Rb1 Qa8 ({I avoided} 26... Nf3+ {because of} 27. gxf3 Bxc3 28. Qb3 ({the computer gives} 28. Nc4 Bd4+ {with perhaps a very faint edge for Black}) 28... Bd4 29. Kf2 {but Black then has} 29... Qd8 30. Qxb7 Qh4+ {and if} 31. Ke2 {then} 31... Qxh2+ 32. Kd3 Qf2 {and apparently Black is winning.}) 27. Qb3 Ba6 {Black has ideas of Nf3+ followed by Bxc3 and g6 followed by Bd3.} 28. Ng4 Nxg4 {This simplification should leave Black with a slight edge.} 29. Bxg4 Bxc3 30. Qxc3 Rb8 31. Rb2 { <span class="PgnWidget-anchor-diagram">[]</span> Here my opponent offered a draw. I felt I was better and had visions of winning a bishop ending with my king on e5 and my bishop on d3, though this may be unrealistic as White can play something like Kf2 and Be2 and the a pawn (helped by the interference of the d pawn) becomes a real danger. We both had about ten minutes plus the fifteen second increment. I still think it&#8217;s right in principle to play on here.} 31... h6 32. h4 g6 33. Kh2 Rb7 (33... Rxb2) 34. Rf2 {I hadn&#8217;t considered this. The game begins to swing to White.} 34... Qe8 35. Bf3 c4 {This felt a bit loose but may be justified.} 36. g3 Rb3 37. Qa5 Qc8 {Not appreciating the danger on the kingside.} (37... Bc8 {and the computer still prefers Black}) 38. Bd1 {This should have let me off the hook.} (38. Bg4 Qa8 39. h5 {gives a dangerous attack.}) 38... Rd3 {Probably losing.} (38... Rb8) (38... Re3 {had been my first thought but it&#8217;s also bad}) 39. Bc2 Re3 (39... c3 40. Bxd3 Bxd3 {is worth a try but} 41. Qa7 Qe8 {looks very good for White}) 40. Qb6 {I think I had missed this.} 40... Re1 {Now I offered a draw} ({It took me some time to realise} 40... Qc5 {loses to} 41. Qd8+ (41. Qxa6 Rxa3 {was the idea}) (41. Qxc5 dxc5 {is hard to assess}) 41... Kg7 42. Qf6+ Kh7 43. Qxf7+ Kh8 44. Qf8+ Kh7 45. Rf7#) 41. Qxd6 {Now White is winning easily.} 41... c3 42. e5 Re2 43. Rxe2 Bxe2 44. e6 fxe6 45. dxe6 Qf8 46. Qxf8+ Kxf8 47. Kg1 Ke7 48. Kf2 Bc4 49. Ke3 g5 50. hxg5 hxg5 51. Kd4 Bxe6 52. Kxc3 Kd6 53. Kd4 Bd7 54. a4 Bc8 55. Bd3 Bh3 56. a5 Kc6 57. Ke5 {and White got the g pawn through.} *
() -
Commented by

In the second, played for Dulwich, my opponent’s draw offer came as a welcome release from suffering, but it may have shown shrewd judgement on his part:

You need to activate javascript to enhance chess game and diagram visualization.
1. d4 {My opponent was delayed by twenty minutes but then moved quickly and confidently, even chatting with his captain for five minutes or so on his own move. I get rather nervous when people move quickly and was wondering if I&#8217;d fallen into some theory he knew.} 1... Nf6 2. c4 c5 3. d5 b5 4. cxb5 a6 5. b6 e6 6. Nc3 exd5 7. Nxd5 Nxd5 8. Qxd5 Nc6 9. e4 (9. Nf3 {is rather more popular and has better results, but this is quite normal.}) 9... Be7 10. Bc4 O-O {All theory so far.} 11. Be3 {This was new to me and turns out to be very rare. The bishop often comes to d2 and c3.} 11... Qxb6 12. Ne2 Qxb2 {Underrating his play. Black will suffer a lot for this pawn.} ({The computer prefers} 12... Bb7 {with some advantage.}) 13. O-O Qe5 {My teammates suggested this was unnecessary and the computer agrees.} (13... Rb8 {is probably better.}) 14. Bf4 {I missed this. The diagonals and b file seem good value for the pawn, though this isn&#8217;t the computer&#8217;s choice.} ({The machine much prefers} 14. Rac1 {e.g.} 14... Qxd5 15. Bxd5 Rb8 16. Bxc5 Rb2 17. Nf4) 14... Qxd5 {Ten minutes&#8217; thought trying to calculate a way to unravel.} 15. Bxd5 { <span class="PgnWidget-anchor-diagram">[]</span> } Bb7 {Perhaps protecting my structure too much.} ({The computer suggests} 15... a5 {with reasonable play.}) 16. Rab1 Ra7 17. Rfd1 ({I had calculated} 17. Bc7 Ba8 18. Bb6 Rb7 {but my opponent continues with sensible moves}) 17... Rc8 {Another ten minutes. I wanted to play d6 to block the diagonal and ultimately contest the b file.} ({Not} 17... d6 18. Rxb7 Rxb7 19. Bxc6) (17... Ba8 {might be better}) 18. Nc3 {I think this was my opponent&#8217;s first long think though this might have been when he was talking to his captain} (18. Bxf7+ {fails to} 18... Kxf7 19. Rxd7 Ba8) 18... Ba8 {Preparing the next move. This took another seven minutes as there is a lot to calculate and my position felt very precarious.} 19. Na4 Nd4 {Gaining a tempo with the threat of Ne2+ and putting the knight on a good square. But this is probably a serious mistake.} (19... Re8) 20. Nb6 {Missing a good chance.} (20. Bxa8 {is met by} 20... Ne2+ (20... Rcxa8 21. Bb8)) (20. Bb8 {can be met by a good exchange sacrifice:} 20... Bxd5 21. Bxa7 Bxe4 (21... Bxa2)) ({But} 20. Rb8 {is strong. The rook exchange reduces the attraction of the exchange sacrifice. A computer-aided line goes} 20... Rf8 21. Rxf8+ Bxf8 22. Bb8 Bxd5 23. Bxa7 Ne2+ 24. Kh1 Bxe4 25. Rxd7 Nf4 26. Nxc5 Bxg2+ 27. Kg1 Bc6 28. Rd8 f6 {and Black is in danger.}) 20... Rd8 21. Kf1 Bxd5 22. Nxd5 Bf8 ({I had wanted to play} 22... d6 {but this loses to} 23. Nxe7+ Rxe7 24. Bxd6) 23. Rb6 d6 24. Rdb1 h6 25. Rb8 Ra8 26. Rxa8 {Other moves also give White enough pressure for the pawn but not much more.} (26. R8b7) (26. Rxd8) 26... Rxa8 27. Rb6 a5 28. a4 (28. Bxd6 Bxd6 29. Rxd6 Rb8 {seems all right for Black.}) 28... Kh7 {Not ideal. I thought Black was almost in zugzwang.} (28... f5 {would have been useful here and at various other points.}) 29. f3 Nc2 ({My teammates were expecting} 29... Rc8 {and pushing the c pawn. I think I was worried I might lose it.}) ({The computer comes up with} 29... f5) 30. Ke2 ({After} 30. Bxd6 Bxd6 31. Rxd6 {the computer suggests} 31... Nd4 {making nonsense of the previous knight move and thinks White is clearly better.}) 30... Nb4 {Time control, which we both made by about 7 minutes. Now twenty minutes&#8217; quickplay.} (30... f5) 31. Nc3 (31. Bxd6 {should give White winning chances as his pieces are closer to the action and Black&#8217;s queenside pawns are shaky.}) (31. Nxb4 axb4 {is very good for Black.}) 31... Rd8 ({I thought about} 31... Nc2 {. I seem to have developed a weakness for irrelevant gestures with my minor pieces.}) 32. Rb7 Kg6 33. Ra7 d5 (33... f5 {is better}) 34. Bc7 Rd7 (34... Re8 {may be better}) 35. exd5 {Giving Black a chance to equalise.} (35. Nxd5) 35... Nxd5 (35... Bd6 36. Bb6 Rxa7 37. Bxa7 Be5 38. Kd2 Bf4+ {sets up a repetition.}) 36. Ra6+ ({the computer prefers} 36. Nxd5 {immediately as Black may need to play f6 anyway.}) 36... f6 37. Nxd5 Rxd5 38. Ra8 Bd6 {The rook ending is the best choice for Black.} (38... Be7 {still offers White some hope.}) 39. Bxd6 ({Not} 39. Rd8 Re5+) (39. Bxa5 {is more complex but Black should be OK}) 39... Rxd6 40. Rxa5 Rd4 { <span class="PgnWidget-anchor-diagram">[]</span> } 41. Rxc5 {Having teetered on the brink of defeat since taking the pawn in the opening, and trailing on the clock, I was relieved when he now took the pawn and offered a draw. I had thought the ending was technically drawn but that White could push for victory. However, a little analysis suggests that it&#8217;s surprisingly easy for White to disappear over the edge if he does so. Maybe my opponent&#8217;s offer was wise rather than merciful.} ({A line that illustrates the problems, with natural though inaccurate play for White, is} 41. Ra8 Rb4 42. a5 Rb2+ 43. Kd3 Rxg2 44. a6 (44. Kc4 {is better}) 44... Rxh2 45. Kc4 Rf2 46. Rb8 Rxf3 47. Rb3 Rf4+ 48. Kb5 ({or} 48. Kxc5 Ra4) 48... Re4 {Just in time.} 49. a7 Re8 {If the king were on c5 White would have Rb8 here.} 50. Rg3+ Kh7 51. Kb6 g5 52. Kb7 h5 53. a8=Q (53. Re3 {is better}) 53... Rxa8 54. Kxa8 h4 {and the pawns will beat the rook.}) *
() -
Commented by

In the third, played for Pimlico, both players were under the illusion that Black could force a draw by repetition. Luckily, before ‘forcing’ the draw, I had another look and saw something better. Actually, this game would be much technically better if I just omitted the repetition altogether; some great players are rumoured to have improved their games for publication. But it’s the story that you dine out on.

You need to activate javascript to enhance chess game and diagram visualization.
1. d4 Nf6 2. c4 c5 3. d5 b5 4. cxb5 a6 5. bxa6 g6 6. Nc3 Bxa6 7. f4 d6 8. Nf3 Bg7 ({Fedorowicz in a similar position recommends} 8... Qa5 {which I considered at some stage, perhaps a move or two later.}) 9. e4 Bxf1 10. Rxf1 Nfd7 {Rare, but not unheard of.} 11. Qe2 Qc7 {Black is trying to prevent e5 but if White plays Nb5 first it should work.} 12. e5 {I expected Nb5 here or on the next move. The next three moves took twenty-five minutes as I needed to calculate.} ({After} 12. Nb5 Qb6 (12... Qb7 13. e5 {and the possibility of Nc7+ protects d5.}) 13. e5 O-O ({I had originally considered} 13... dxe5 14. fxe5 Nxe5 15. Nxe5 Bxe5 16. Qxe5 {but then realised the rook was en prise at the end.}) 14. a4 {White is doing well} (14. exd6 exd6 {opens the file too early}) (14. e6 fxe6 15. Qxe6+ Kh8 {worried me until I realised the knight was en prise e.g.} 16. Ng5 Qxb5)) 12... dxe5 13. fxe5 (13. Nb5 {is like the previous note}) 13... Nxe5 14. d6 {Another surprise, and this time a mistake.} ({I was looking at something like} 14. Nb5 Nxf3+ 15. Rxf3 Qd7 16. d6 O-O) 14... Qxd6 15. Nb5 Nd3+ {The right move, though for the wrong reason} 16. Kd2 { <span class="PgnWidget-anchor-diagram">[]</span> } Bh6+ {?? As the game goes, this is a superfluous repetition, but it should have cost the game.} (16... Qf4+ {gives the game continuation two moves later}) 17. Kc3 {?? Originally I had missed this and then saw I could repeat the position.} ({We both seem to have missed that after} 17. Kc2 {(which was the main move I had considered)} 17... Nb4+ 18. Kb1 {Black&#8217;s knight has escaped but the bishop is attacked and he cannot save both his attacked pieces e.g.} 18... Qd7 {angling for Qf5 +} 19. Ne5 ({or} 19. Bxh6)) 17... Bg7+ 18. Kd2 (18. Kb3 Qd5+ 19. Kc2 Nb4+ ({rather than my planned} 19... Qc4+ {when White staggers on with} 20. Nc3) 20. Kb1 Qf5+ {forces mate.}) ({if} 18. Kc4 Ra4+ 19. Kb3 {the computer finds} 19... Nxc1+ {and if} 20. Rfxc1 {then} 20... Rb4+ 21. Ka3 Qa6#) {Before repeating again I thought I should check whether I had something better and realised I did, though not how bad Bh6+ was.} 18... Qf4+ {At about this point I thought I heard a very quiet &#8216;oh shit&#8217; from my opponent&#8217;s side of the board. I wish my opponents would do this more often; it helps with evaluation.} 19. Kxd3 ({My opponent wondered afterwards if he could have tried} 19. Qe3 {but I don&#8217;t think it helps much}) 19... c4+ 20. Kc2 Qf5+ 21. Kd1 Qxb5 {My opponent had missed that this knight was dropping. I now wrongly thought I was two pawns up; it&#8217;s only one, but in any case Black is much better. I felt I should have put him away more expeditiously, but my opponent manages to stabilise his disadvantage.} 22. Ke1 Nc6 23. Kf2 O-O 24. Kg1 e6 (24... Nd4 {is possible}) 25. Kh1 Nd4 26. Nxd4 Bxd4 27. Bh6 Qxb2 28. Qxc4 Rac8 (28... Rfc8 {is also possible}) 29. Qa6 Bg7 ({The computer greatly prefers} 29... Qxa1 30. Rxa1 ({not} 30. Bxf8 Rc1) 30... Bxa1 {when the two rooks will probably corral the a pawn easily enough. I think I had the vague feeling that I wasn&#8217;t sure how to handle two rooks against a queen.}) ({I tried to find something with} 29... Rc2 30. Bxf8 Rxg2 {but fortunately I gave the idea up.} 31. Bd6 {is devastating.}) 30. Rab1 Qe5 31. Bxg7 Kxg7 { <span class="PgnWidget-anchor-diagram">[]</span> I am not absolutely sure if Black is now winning. With this pawn formation Black should win a pure rook endgame but not a pure queen endgame. Ultimately Black wants to combine threats of a queen exchange with threats to the king and the loose a pawn, but White generates some tricky counterplay.} 32. Rb7 Rc7 33. a4 Rfc8 34. h3 (34. Qa7 {might be worth a punt as Black must avoid} 34... Qe2 35. Rxf7+) 34... Rxb7 35. Qxb7 Rc7 36. Qb8 Qc3 37. Qb6 Qc5 38. Qb2+ Qc3 39. Qb6 {About here my opponent, with only five minutes left, offered a draw. I had about eight minutes and felt I should win. I got this one right, anyway} 39... Qc4 40. Rb1 Rc6 41. Qb2+ Qc3 42. Qb5 Rc5 43. Qb7 Qc4 44. Qb6 Qd5 45. Rf1 Rc2 46. Rg1 (46. Qg1 {is rather better}) 46... Ra2 47. Qb4 e5 48. Qg4 Qd4 49. Re1 Qxg4 (49... Rxa4 {wins another pawn but I think the rook ending is winning}) 50. hxg4 f6 (50... Rxa4 51. Rxe5 Rxg4 {might be simpler}) 51. Re4 Kf7 52. Kh2 Ra3 53. g3 Ra2+ 54. Kh3 {Odysseus in slippers.} (54. Kg1 {must be better}) 54... Ke6 55. Rb4 Kd5 56. Rb7 e4 57. Rxh7 e3 ({Keeping the white king incarcerated rather than taking the a pawn. I wondered about using the king to deal with the a pawn first by} 57... Kc5 {which probably also works but is less clear.}) 58. g5 f5 59. Re7 e2 {This does not release White&#8217;s king because of the discovered check along the second rank.} 60. a5 Kd4 61. a6 Kd3 {Now when a7 is played White won&#8217;t be in a position to play Rxe2.} 62. Rd7+ ({With more time White might have tried} 62. a7 {when I just might have gone for} 62... Kd2 (62... Rxa7 {gives an easy win with queen versus rook, but I wanted to win his rook}) 63. Rd7+ Ke1 {??} (63... Ke3 {wins}) 64. Kg2 {and White holds. This line literally stopped me in my tracks when it occurred to me on the way home.}) 62... Ke3 63. Re7+ Kf2 64. g4 fxg4+ (64... f4 {is aesthetically preferable but I may have seen a ghost of stalemate.}) 65. Kxg4 {Here his clock was down to one second before the increment kicked in.} 65... e1=Q 66. Rxe1 Kxe1 {and White resigned.} *
() -
Commented by

 

 

 

Brief lives

Miniature chess games are usually considered to be anything under twenty-five moves or so. They include, of course, many great masterpieces. Games under ten moves are rather different. I am not sure if anyone has done an anthology: there is surely scope. There is an anthology of very short poems, which restricted the length to thirteen lines (I think) to exclude sonnets, and of course there are many anthologies of haiku, tankas, clerihews, epigrams and so on. But the perfection of a very short and decisive game of chess rests essentially on the absoluteness of the error it contains; like a Persian carpet where the pattern occurs only once and the obligatory flaw must occur in its only occurrence.

There is a familiar trap in the Englund gambit which I have fallen into more than once online:

You need to activate javascript to enhance chess game and diagram visualization.
1. d4 e5 2. dxe5 Nc6 3. Nf3 Qe7 4. Bf4 Qb4+ 5. Bd2 (5. Qd2 {loses to} 5... Qxb2 6. Qc3 Bb4) 5... Qxb2 6. Bc3 (6. Nc3 {is the remedy when} 6... Nb4 7. Nd4 c5 8. Rb1 {wins for White e.g.} 8... Qa3 9. Ndb5 Qa5 10. a3 Nc6 11. Nd5) 6... Bb4 7. Qd2 Bxc3 8. Qxc3 Qc1# 0-1
() -
Commented by

Occasionally, I get lucky online and win very quickly. The English contains a number of positions that are sharper than they look. This blitz game was played at 3 minutes with a two second increment.

You need to activate javascript to enhance chess game and diagram visualization.
1. c4 e5 2. Nc3 g6 {a bit premature} 3. d4 {sniffing the weak diagonal} 3... exd4 (3... Bg7 {may be best}) 4. Qxd4 Nf6 { <span class="PgnWidget-anchor-diagram">[]</span> already losing} (4... f6 {and White has a nice game}) ({I discover John Watson discusses} 4... Qf6 {and suggests} 5. Qe3+ Qe6 6. Nd5 Qxe3 7. Bxe3 Na6 8. Bd4 f6 9. O-O-O Kf7 10. Nf3 c6 11. Nf4) 5. Bg5 Be7 (5... Bg7 6. Nd5 {is similar though it would avoid the specific trick in the game}) 6. Nd5 Nc6 7. Nxf6+ ({winning though the machine prefers} 7. Qc3) 7... Kf8 {maybe setting a trap} (7... Bxf6 8. Qxf6 Qxf6 9. Bxf6 O-O {and White is simply a piece up}) 8. Bh6# ({Even} 8. Qc3 {is better than it looks after} 8... Bb4 9. Bh6+ Ke7 10. Nd5+) 1-0
() -
Commented by

Such games offer the winner an odd mixture of aesthetic pleasure and a facile sense of achievement. They are rare in competitive chess, but they do happen. At the end of last season, I turned up to a match for Dulwich in the London League exhausted after some building work at home and found myself playing Black against a strong player. I have never felt more like offering an early draw. It turned out there was no need. The following took less than ten minutes:

You need to activate javascript to enhance chess game and diagram visualization.
1. d4 Nf6 2. c4 c5 3. d5 b5 4. Nf3 {One of the surprises about playing the Benko is how few people accept it.} 4... e6 {Transposing into the Blumenfeld.} (4... Bb7 {is also good.}) 5. Bg5 (5. dxe6 {is also possible and was Karpov&#8217;s choice when Lobron played the Blumenfeld against him, though Black&#8217;s central majority can be dangerous. Alekhine overwhelmed the elderly Tarrasch in an early game. But I am not yet sure I&#8217;ve got the hang of the Blumenfeld.}) 5... exd5 { <span class="PgnWidget-anchor-diagram">[]</span> Now my opponent started to think, not always a bad idea but one with some risks. I wondered what he was considering and amused myself with a fantasy line where he got mated in a few moves.} 6. Bxf6 (6. cxd5 d6 {is the main line of the Blumenfeld. White can often get some edge by gaining control of c4.}) 6... Qxf6 7. Qxd5 {I began to realise my opponent had gone seriously off-piste. The main technical challenge now is keeping a straight face.} ({The computer suggestion} 7. Nc3 {generates interesting play for White where natural play for Black can be dangerous e.g.} 7... d4 (7... dxc4 {may give Black some advantage}) 8. Nd5 Qd6 9. e4 dxe3 ({better} 9... Bb7) 10. cxb5 exf2+ (10... Bb7 {is necessary}) 11. Kxf2 {and White&#8217;s quicker development gives a decisive advantage e.g.} 11... Be7 12. Bc4 O-O 13. Nxe7+ Qxe7 14. Qd5) 7... Qxb2 8. Qxa8 ({White can bale out here with} 8. Qe5+ {but there is no way a strong player would aim for White&#8217;s position after} 8... Qxe5 9. Nxe5 d6 (9... g5 {is a computer improvement}) 10. Nf3 bxc4 {so I was not altogether surprised by his choice.}) (8. cxb5 Qxa1 9. Nd2 Nc6 {is another computer suggestion, also very bad for White but a little more complicated.}) 8... Qc1# {I simply played this and stopped the clock. My opponent took it very well; these things happen though not usually on move 8.} ({I imagine he had analysed} 8... Qxa1 {though the computer prefers Black here too.}) 0-1
() -
Commented by