Jon Speelman's Agony Column #41

[Event "Coventry League"] [Site "?"] [Date "2006.02.14"] [Round "?"] [White "McConnell, Peter"] [Black "Jones, Michael"] [Result "1-0"] [ECO "D05"] [Annotator "Jones,Michael"] [PlyCount "39"] [EventDate "2006.??.??"] [SourceDate "2015.07.13"] 1. d4 Nf6 2. Nf3 e6 3. e3 d5 4. Bd3 c5 5. c3 Nc6 6. Nbd2 Be7 7. O-O O-O 8. dxc5 Bxc5 9. e4 dxe4 10. Nxe4 Nxe4 11. Bxe4 Qc7 $2 {[#] Completely overlooking the danger. JS The Greek gift sacrifice is one of the main points of the Colle System, and this position has been played several times before. In fact it turns out that the sacrifice should only lead to a draw.} 12. Bxh7+ $1 { JS I guess that really this deserves a !? rather than an !} Kxh7 13. Ng5+ Kg6 $1 {[#]} ({Marching the king up the board doesn't look great, but the alternative is horrible JS You often have to play Kg6 or Kh6 (obviously not that here with the bishop on c1) after the sacrifice and in fact Black should be perfectly okay. Indeed after White's next move he should win!} 13... Kg8 14. Qh5 {JS Yes indeed. Michael was absolutely right to avoid this. The line continues:} Rd8 (14... Re8 15. Qh7+ Kf8 16. Qh8+ Ke7 17. Qxg7 Ne5 18. Ne4 { is equally bad}) 15. Qh7+ Kf8 16. Qh8+ Ke7 17. Qxg7 Ne5 {JS and obviously White has a huge attack, for instance} (17... Rf8 18. Nh7) 18. Ne4 Kd7 19. Nxc5+ Qxc5 20. Rd1+ Kc7 21. Bf4) 14. h4 $2 {Black isn't completely lost here, but has to find the correct defence. I didn't. JS After the shock of Bxh7+, Michael is underestimating his chances.} (14. Qd3+ $2 f5 15. Qh3 Be7 (15... f4 $1 {JS so that the bishop doesn't protect the knight}) 16. Qh7+ Kf6 17. Qh5 Bd6 18. Nh7+ Ke7 19. Bg5+ Rf6 20. Qg6 Kd8 21. Nxf6 gxf6 22. Bxf6+ Kd7 23. Rad1 Ne7 24. Qf7 Kc6 25. Rxd6+ Kxd6 26. Bxe7+ Kc6 27. Rd1 b6 28. Qe8+ Kb7 29. h4 Ka6 30. h5 Qf4 31. Qc6 Bb7 32. Qxe6 Qg4 33. f3 Bxf3 34. Rd2 Rg8 35. Qd6 Bxg2 {1/2-1/2 (35) Schroeder,S (1845)-Fernandez-Alvarez,R (1723) Bayerisch Eisenstein GER 2014}) (14. Qg4 f5 15. Qh4 {[#]} Bd6 (15... f4 $5 16. g4 $3 {is the difference from the queen being on h3} fxg3 (16... Bd7 17. Qh5+ Kf6 18. Ne4+ Ke7 19. Qxc5+ ) 17. hxg3 {and the attack is still very strong} Qe5 18. Qh7+ Kf6 19. Bf4 Qf5 20. Ne4+ Kf7 21. Qxf5+ exf5 22. Nxc5 {seems to be the best Black can get but is a fairly mild disadvantage, so perhaps the most sensible even if Bd6 is better "in theory".}) (15... Bd7 16. Qh7+ Kf6 17. Qh5 Ke7 18. Nh7 $1 (18. Ne4 fxe4 19. Bg5+ Rf6 20. Bxf6+ gxf6 21. Qxc5+ Qd6) 18... Ne5 19. Bg5+ Rf6 20. Rad1 Rh8 21. Qh4) (15... Be7 16. Qh7+ Kf6 17. Qh5 g6 18. Qh4 Bd8) 16. Rd1 (16. Qh7+ Kf6 17. Qh5 Ke7) 16... Bd7 17. Qh7+ Kf6 18. Qh4 Kg6 19. Qh7+ Kf6 20. Qh4 { with a draw unless Black wants to try his luck with Be7 since} Rh8 $2 {loses to } 21. Ne4+ Kf7 22. Nxd6+ Qxd6 23. Rxd6 Rxh4 24. Rxd7+) 14... f6 $4 {Rushing headlong towards my doom. With the knight next to my king, my instinct was to kick it away to gain f7 as a flight square, but I failed to see that White is saved the effort of having to move it by the existence of a forced mate.} ( 14... f5 {makes f6 available as an escape route instead, after which Black is only slightly worse.}) (14... Rh8 $1 {JS leaves Black a piece up for really not very much since if} 15. Qg4 ({or} 15. Ne4 Rxh4) 15... f5 16. Nxe6+ fxg4 17. Nxc7 g3 $1) 15. h5+ Kf5 16. Qf3+ Ke5 17. Bf4+ Kf5 18. Bxc7+ ({With the queen en prise White goes for that first, missing the immediate} 18. g4# {but it only prolongs the agony by a couple of moves.}) 18... Kxg5 19. Bf4+ Kf5 20. g4# {[#] A St. Valentine's Day massacre.} 1-0

[Event "Coventry League"] [Site "Divisional Cup"] [Date "2010.02.09"] [Round "?"] [White "Paterson, Andrew"] [Black "Jones, Michael"] [Result "0-1"] [ECO "B03"] [Annotator "Jones,Michael"] [PlyCount "116"] [EventDate "2010.??.??"] [SourceDate "2015.07.13"] {Andrew was something of an unknown quantity, having recently returned to competitive play after several years' absence, but hadn't shown any signs of rust in winning his first three games, including grinding me down in a rook endgame the previous week. In contrast, I had a solitary half point from three games since Christmas.} 1. e4 Nf6 {The Alekhine is a weapon I bring out occasionally, mostly in an attempt to avoid getting dragged into someone's pet line in a more common opening. I've had some good results with it, but also seen a few times when my opponent turned out to know it better than me, and I got completely mauled.} 2. e5 Nd5 3. d4 d6 4. c4 Nb6 5. exd6 cxd6 6. Nc3 Bf5 7. Nf3 Nc6 8. Bd3 {This surprised me slightly - it exchanges one of Black's bishops while the other is undeveloped, but also allows me to kick the white queen and threaten a fork on c2} Bxd3 (8... Bg4 {JS isn't bad}) 9. Qxd3 Nb4 10. Qe2 {[#]} Nxc4 {I see a combination and dive in. JS This is extremely optimistic and should lead to very serious trouble.} 11. Qxc4 $5 {I was surprised Andrew allowed the fork. Presumably he thought that once the knight reached a1 it would not escape from there, so he would end up with two knights for the rook.} (11. a3 {JS was even stronger} Qa5 12. Qxc4 Nc2+ 13. Ke2 Nxa1 14. Nb5 Kd8 (14... Kd7 15. Ne5+ $1 {as the software goes ping!}) 15. Bd2 Qb6 16. Rxa1 Rc8 17. Qa4) 11... Nc2+ 12. Kd1 (12. Ke2 Nxa1 13. Be3 {JS and White will easily win the knight with a clear advantage at the very least.} Rc8 ( 13... Qb6 14. Nd5 (14. Qa4+ Qc6 15. Rxa1 {JS is simple}) 14... Qxb2+ 15. Nd2 Qc2 {JS and White has various ways to get what should be a winning advantage for instance} 16. Nc7+ Kd8 {This way rather than Kd7. There isn't Rc7+ later on but actually that isn't a serious worry anyway.} 17. Nxa8 Qxc4+ 18. Nxc4 Nc2 19. Rc1 Nxe3 (19... Nb4 20. Rb1) 20. Nxe3) 14. Qb5+ Qd7 15. Qxd7+ Kxd7 16. Rxa1 ) 12... Nxa1 13. Bf4 {JS An odd square for the bishop} (13. Be3 {was more normal} Rc8 14. Qb4 Qd7 15. Kd2 {with a nice advantage.}) 13... Qd7 {[#] Putting the queen on a useful diagonal.} 14. Nd2 $2 {JS This discoordinates White} (14. Kd2 {was correct}) 14... Qg4+ {The hanging bishop forces the reply} 15. Ne2 Rc8 {The point of clearing the c-file: the knight will now be able to escape via c2, after which Black is just an exchange up.} (15... Qxg2 {is also fine. After the rook moves, the white queen will come to c1 and pick off the knight in the corner, but Black will have more than adequate compensation. If} 16. Rg1 Qc6 {to prevent Qc1,} 17. Qxc6+ bxc6 18. Kc1 {the knight will fall, but Black still has rook and two pawns for two knights.}) 16. Qb5+ Kd8 (16... Qd7 $1 17. Qxd7+ Kxd7 18. Nc3 {JS and even if White could round the knight up it would be after a huge effort giving Black time to cause serious problems} ( 18. Nf3 Nc2 19. a3 (19. Bd2 b5 20. Ng5 h6 21. Nxf7 Rh7) 19... Na1 $1 {JS as Houdini has just told me and the knight escapes}) 18... b5 (18... g6 19. Nf3 b5 $1) 19. Nxb5 Rb8 20. Nc3 Rxb2) 17. f3 Qg6 (17... Qd7 {JS was still correct. Since Black will have huge trouble getting the a1 knight out he will be in trouble}) 18. Ne4 {[#]} Qxg2 $4 {A case of "chess blindness". I had refrained from playing Qxg2 on the previous move, seeing the combination which would follow, but somehow I forgot about that and played it on this move instead.} 19. Rg1 Qxf3 20. Ng5 Qc6 21. Qxc6 Rxc6 22. Nxf7+ Ke8 23. Nxh8 {A position which wouldn't look out of place in a game between complete beginners. Both sides have allowed a knight fork, and both now have a knight in the corner. White has a piece for two pawns and is clearly better, but things aren't over yet, particularly since my knight can come to c2immediately, while his isn't going anywhere for the time being.} Nc2 24. a3 e5 {With the intention of deflecting the bishop from the defence of e3 in order to extract the knight, but not good.} ({The immediate} 24... g6 {was best. After} 25. d5 Rc5 {Black regains the piece and is close to equality. JS This is a position where the lines go on for quite a while as White tries to find ways to exploit the knight.} 26. Nxg6 hxg6 27. Rxg6 Kf7 28. Rg3 Bg7 29. Rf3 Ke8 30. Rd3 {is a sensible start and it does indeed seem to be approximately equal.}) 25. dxe5 dxe5 26. Bxe5 {Now if 26...g6, the knight is immediately defended and there is no ...Bg7 winning it.} Ne3+ 27. Ke1 Nf5 28. Rf1 g6 29. Nd4 {JS Extremely natural though my engine prefers} (29. Rf3) 29... Rc1+ {The point of exchanging rooks is to prevent White defending f7, keeping the knight in the corner.} 30. Ke2 Rxf1 31. Kxf1 Bg7 32. Bxg7 Nxg7 {Now for a succession of errors from both sides, as White manoeuvres his other knight in an attempt to defend f7 and extricate the one in the corner, Black aims to attack it before it can get out but neither finds the best way of doing so. JS White shouldn't be able to save the trapped knight in the long term, but can wreak havov with the other knight, and this should win.} 33. Nb5 g5 $2 ({With the idea that when the knight is trapped, it won't be able to grab a pawn before being lost. But this is misguided. d6 is wide open and the immediate} 33... Nf5 {is necessary.} 34. Nxa7 Kf8 35. Nb5 Kg7 36. Nxg6 hxg6 {and although White should win, there's still some work required to achieve it. JS A very sensible assessment by Michael. The most obvious line is very close, so White would have to manoeuvre:} 37. Ke2 Kf6 38. Kd3 (38. Kf3 $1 {is better}) 38... Ke5 39. Kc4 Ne3+ 40. Kc5 g5 41. a4 g4 42. a5 Nf1 43. Nc3 Nxh2 44. Ne2 Ke4 45. Kb6 Kf3 46. Ng1+ Kf2 47. Kxb7 Kxg1 48. a6 g3 49. a7 g2 50. a8=Q Nf3 {[#] and Black holds since he will queen and be able to stop the b pawn after the queens are exchanged, for instance} 51. b4 Kf2 52. Qa7+ Kf1 53. Qe3 Ne1 54. b5 g1=Q 55. Qxg1+ Kxg1 56. b6 Nd3 57. Kc6 {[#]} Ne5+ $1 58. Kc7 Nd3 $1 59. b7 Nc5) 34. Nd6+ Kf8 35. Ke2 ({White isn't forced to move the knight immediately, but there's no particular reason not to do so. After} 35. Nhf7 Ke7 {neither knight can move for the time being, but neither can be captured either, so White has plenty of time to bring in the king to defend one knight, enabling the other to move.}) 35... Ne6 {[#]} 36. Ke3 $4 {White thinks he has one more tempo to remove the knight from h8 than he actually does. Now the immediate 36...Kg7 fails to 37.Nhf7 but instead...} (36. Nhf7 Ke7 37. Kf3 h5 38. Ke4 g4 39. Ke5 h4 40. Nh6) 36... Nd8 $1 {Whoops: now the knight cannot escape. JS A splendidly dirty trick even though White is still much better.} 37. Ke4 Kg7 38. Kf5 Kxh8 39. Kf6 ({I couldn't understand why White didn't just play} 39. Kxg5 {here, when he has a slight edge thanks to his more active king and knight, but I should just about be able to hold. JS Kf6 looks appealing, but indeed simply Kxg5 was stronger.}) 39... g4 40. Nf5 {Removing the attack on b7 allows me to activate my knight.} Nc6 41. Nh6 Na5 42. b4 (42. Kf7 Nc4 43. Kf8 $2 Ne5 { JS and White is in zugzwang} 44. b3 b5 45. b4 a6) 42... Nc4 43. Nxg4 Nxa3 ({ The zwischenzug} 43... h5 {was better: after} 44. Ne5 Nxa3 {Black retains an extra pawn, albeit with a drawish position, and avoids the unpleasantness I ran into a few moves later.}) 44. Nh6 Nc2 45. Kf7 {At this point I momentarily froze, noticing the threat of 45...-- 46.Kf8 -- 47. Nf7# Fortunately I can bring the knight back just in time to prevent this.} Nd4 (45... Nxb4 $4 46. Kf8 {and mate next move.}) 46. Kf8 Ne6+ 47. Kf7 Nd8+ 48. Kf8 b6 $2 {Aiming to create a passed pawn on the queenside, but this is not the way to do it.} ( 48... a6 {had to come first. JS Yes, you must (almost - there must be an exception) always avoid backward pawns when mobilising a majority} 49. Nf5 b6 50. Nd4 a5 51. b5 Nb7 52. Ke7 {JS White is active enough here to draw, but certainly no more.} Kg7 53. Kd7 Nc5+ 54. Kc6 Na4) 49. Nf5 {White had two ways to ensure a draw here, and unfortunately for him this isn't one of them. It doesn't lose immediately, but does make his task rather more difficult.} (49. Ke7 {By kicking the knight away, White can renew the mate threat and aim for a repetition.} Nc6+ ({Black can choose to avoid the repetition by} 49... Kg7 { but this still draws:} 50. Kxd8 a5 (50... Kxh6 51. b5 Kh5 52. Kc7 Kh4 53. Kb7 Kh3 54. Kxa7 Kxh2 55. Kxb6 h5 56. Kc7 h4 57. b6 h3 58. b7 Kg1 59. b8=Q h2 { is a book draw.}) 51. bxa5 bxa5 {and the knight must transfer to the queenside to stop the a-pawn, leaving the black king to pick off the white h-pawn.} 52. Nf5+ Kf6 53. Nd4 a4 54. Nc2 Kg5 55. Kc7 Kh4 56. Kb6 Kh3 57. Kb5 Kxh2 58. Kxa4) 50. Kf8 Nd8 51. Ke7 Nc6+ 52. Kf8) ({While the quiet} 49. b5 {blocks the pawns and also secures the draw. JS This is much the simplest} Ne6+ 50. Kf7 Nd4 51. Kf8 {and the threat of Nf7# ensures that Black will never have time to capture on b5.}) 49... a5 50. bxa5 bxa5 {[#]} 51. h4 $2 {Fiddling while Rome burns: White's h-pawn is never going to promote, he needs to concentrate on stopping the black a-pawn first. JS Yes this is the losing move.} (51. Nd4 {and Black may still be able to win, but will have a much tougher task. JS Assessments by software are often useless (to humans at least), but here I trust an engine to be able to anlayse far enough ahead to get it right. I'm getting just -0.6 odd here, so I presume that White can hold. One line goes} a4 52. Ke7 $1 Nb7 53. Nc2 Kg7 54. Kd7 Na5 55. Kd6 Nb3 56. Kc6 Nd2 57. Kb5 Nf3 58. Kxa4 Nxh2 59. Kb5 h5 60. Ne3 Ng4 61. Ng2 Kf6 62. Kc4 Kf5 63. Kd3 {and White is clearly now drawing easily}) 51... a4 52. h5 ({There's still time to delay the a-pawn's march with} 52. Nd4 {JS But after} a3 53. Nc2 a2 {the pawn is too far advanced. White will have to spend six moves with his king to capture it, and by then the black h-pawn will be far too far advanced.}) 52... a3 53. Ne3 ({Now it's too late for the alternative attempt to stop the pawn, as} 53. Nd4 {meets with the flashy response} Ne6+ $3 54. Nxe6 a2 {when White can still set a nasty trap with} 55. Nd8 {but Black can sidestep it by} (55. h6 {JS} a1=Q 56. Nd8 { fails to} Qa3+ {or} (56... Qa8)) 55... h6 ({Not the unfortunate} 55... a1=Q $4 56. Nf7#)) 53... Ne6+ 54. Kf7 Nd4 {With c2 defended, there's no way of stopping the a-pawn.} 55. Kf8 h6 ({Not strictly necessary, since the immediate } 55... a2 {also wins. But when I'd spent a fair amount of time over the last few moves worrying about the mate threat in the corner I thought it wouldn't do any harm to give the king an escape square.}) 56. Ng2 a2 57. Nf4 a1=Q 58. Ng6+ Kh7 {There have been plenty of tricks on both sides up to this point, but nothing White can conjure up is going to save him now. A demonstration, if anything, of how badly it's possible to play and still win. As if the game itself wasn't enough for me to feel sheepish about, the other three boards were drawn so this comedy of errors decided the match. JS I'm definitely not against swindling. The world's best players all have an excellent "swindle mode" when required!} 0-1