Commentary: Tashkent FIDE Women's Grand Prix 2013, Round 9 (Xue - Lagno)

This next commentary is from round 9 of the women's FIDE Grand Prix in Tashkent.  The game caught my eye because of the unusual exchange sacrifice offered by White (but never accepted) and some of the thematic moves and "little tactics" that led Zhao Xue to win this game in the Symmetrical English against Kateryna Lagno.  Well worth the time to analyze.

[Event "FIDE WGP Tashkent"] [Site "Tashkent UZB"] [Date "2013.09.28"] [Round "9.1"] [White "Zhao, Xue"] [Black "Lagno, Kateryna"] [Result "1-0"] [ECO "A37"] [WhiteElo "2579"] [BlackElo "2532"] [Annotator "ChessAdmin/Houdini"] [PlyCount "85"] [EventDate "2013.09.18"] 1. c4 g6 2. Nc3 c5 {Black heads for a Symmetrical English variation. Other possibilities here would include a KID or Gruenfeld-type setup.} 3. Nf3 Bg7 4. g3 Nc6 5. Bg2 e6 6. O-O {White chooses the most common move, which is a solid line; d4 here would lead to a more active line with a pawn sacrifice. Interestingly, White only scores about 47 percent with the text move.} Nge7 7. a3 {both prophylactic against ...Nb4 and prepares an eventual b4 push. With this variation, a less frequently played one, White scores over 51 percent.} d5 {Black decides to commit in the center, rather than the more popular O-O.} 8. d3 O-O (8... dxc4 9. dxc4 Qxd1 10. Rxd1 Bxc3 11. bxc3 {is something a Class player might prefer, exchanging off as many pieces as possible with Black and inflicting doubled c-pawns on White. In concrete terms, however, White is left with the two bishops and the dark-square one is a monster, with Black's dark-square weaknesses beckoning. Play could continue} h6 12. Bf4 b6 13. Bd6 Bb7 14. Ne5 {and Black is already under significant pressure.}) 9. Bg5 { a thematic idea, designed to provoke a weakening Black pawn advance.} h6 10. cxd5 exd5 (10... hxg5 11. dxc6 Nxc6 {looks fine for White.}) 11. Bxe7 {the decision to exchange the dark-square bishop involves a tradeoff of time (in this case gained by avoiding a retreat) versus the two bishops, which Black now gains.} Nxe7 12. Rc1 Bd7 {this doesn't seem to be the best placement for the bishop. Playing it to e6 or eventually to b7 would keep it more active.} 13. b4 cxb4 {Black chooses to enter an IQP structure in exchange for greater piece activity.} 14. axb4 Qb6 $11 (14... a5 $15 {is what Houdini prefers here, immediately looking to create a passed pawn.} 15. bxa5 Qxa5) 15. Qb3 a5 16. Nxd5 {this tactical option was not possible in the previous variation, where the queen was on a5 instead of the more vulnerable b6 square.} Nxd5 17. Qxd5 Bc6 (17... Qe6 $5) 18. bxa5 $14 {another tactical follow-up exploiting the location of the Qb6.} Qxf2+ 19. Kxf2 Bxd5 {White is a pawn up here, with the passed d-pawn. Black has some compensation in terms of the two bishops and White's potentially weaker pawns.} 20. Rc5 Rfd8 21. Rb1 Bf8 22. Rcb5 Bc6 23. Ne5 {an excellent idea from Zhao and confirmed by Houdini. The exchange sacrifice is dangerous for Black to take.} Rdc8 (23... Bxb5 24. Rxb5 {the b-pawn will now fall as well.} Rac8 25. Rxb7 $16) 24. d4 $16 {passed pawns must be pushed!} Bg7 {based on the game continuation, this seems to be a loss of time, as the bishop will move back after a few moves.} (24... Bxb5 {the exchange sacrifice is still good for White, but at least Black would get rid of one of the powerful rooks on the queenside.} 25. Rxb5 Ra6 26. Bxb7 Rf6+ 27. Bf3) 25. Bxc6 bxc6 26. Rc5 {White takes advantage of the absence of the Bf8.} Ra6 27. Nf3 {choosing to consolidate rather than play more aggressively with something like Rb7.} Bf8 28. Rb6 {what masters call a "little tactic".} Rca8 ( 28... Bxc5 29. Rxa6 Bd6 $18 {and now it will be much easier for White to try and push through the passed a-pawn, whose queening square cannot be covered by the bishop.}) 29. Rcxc6 Rxa5 {although material is still even after the tactic, now nothing stands in the way of the d-pawn.} 30. Rb7 Ra2 31. Rcb6 Re8 32. Rb2 Raa8 33. Rc2 Bd6 {again a bishop move appears to lose time, as it could have moved directly to g7 instead of taking three tempi to do it per the game continuation. The idea appears to be that the bishop covers the e5 square, which it could also do from g7.} 34. Rd7 Ra6 35. Rb2 Bf8 36. Ne5 Bg7 {the f-pawn is protected tactically.} 37. Nf3 {given the lack of progress, one must think that White (and perhaps Black too) was looking to make a series of safe moves in order to save some time on the clock.} (37. Nxf7 $2 Rf6+) 37... Rae6 38. Rc2 Re4 39. Rd2 Bf8 $2 {this bishop is now truly a drag on Black's play. White immediately seizes the opportunity on e5, left uncovered by the bishop.} (39... g5 $5 {Houdini thinks that Black's best option is to try to generate counterplay on the kingside.}) 40. Ne5 $18 Bc5 {a desperation move.} (40... Bg7 $2 {no longer helps, now that Black's Re4 is cut off from assistance and vulnerable.} 41. Kf3 {and Black loses material.}) 41. Nxf7 (41. dxc5 {is also winning for White.} R4xe5 42. c6 Rc5 43. c7) 41... Bxd4+ 42. R2xd4 (42. R7xd4 $2 {White has no need to abandon the knight, as the e-pawn can be lost without damaging her winning prospects.} Kxf7 43. Rxe4 Rxe4 $14) 42... Rxe2+ 43. Kg1 { and now White is a full piece up and her king cannot be trapped.} 1-0

How Irina Krush makes us feel better about chess

From GM-elect Irina Krush's Chess Life Online article about the Baku Open, where she gained the Grandmaster title.

Maybe the biggest [reason] is that despite not having a lot of confidence, for the most part I play like I do. I know my opponent's rating, but it's just a number, not my sentence. For me, chess is a fight, sixty four squares where you lay out everything you have, and I believe in my ability to fight, because it's really just a function of your ability to give everything you have, to put it banally, 'to do your best.' I want to make the maximum effort, whether that means pushing myself to find the best moves, being resilient in defense, or overcoming any psychological weakness that can come up during a game: inclinations towards cowardice, towards giving up in difficult positions, or slacking off in better ones. So while I just can't see myself to be very good in the actual playing of chess, I do come into every game with the belief that I can give it 100%, and that's probably not a lot less than what my opponents can bring. That's where my confidence comes from :)

 

Chess performance and chess skills: not the same thing

The last two analyzed games (Annotated Game #104 and Annotated Game #105) provided a useful reminder that chess performance - by that, I mean a player's actual results in their games - is not necessarily a reflection of a player's chess skills.  This will not surprise longtime players, but it is still worth examining as a phenomenon.

It is said that there is no luck in chess, but that is not quite true.  Because it is a mental game (or sport), anything that affects our mental abilities on a particular day can significantly influence our performance at the board.  Life events, personal issues, random encounters and of course our physical health can all perturb our minds.  I recall that during one tournament, I drove about ten minutes away in order to pick up lunch and bring it back, with not a lot of time to spare before the next round.  Of course that was the time a police officer decided to pull me over for "aggressive driving" because I kept trying to beat the series of lights on the highway which were badly timed.  I hadn't in fact done anything wrong (wasn't speeding and hadn't run a light), so nothing happened to me, but I didn't have the best mental focus for the next round (also having to eat a cold lunch by that point).

Because your skills aren't going to improve while playing in a tournament (or an individual game), what therefore matters most about your performance is maximizing the use of what skills you already possess.  This requires concentration, focus and good judgment.  The flip side of this, of course, is that if you can disrupt your opponent's focus and judgment, your own relative winning chances increase.  I don't advocate intentionally annoying your opponent off the chessboard, but rather doing everything possible to disrupt their play on the chessboard.  This is especially needful when in objective terms you are losing.

As the analysis showed in the last two games, I was either objectively lost or at a serious material and positional disadvantage.  This indicates a failure of chess skills (thinking process, calculation, prophylaxis, etc.) but performance-wise I still won both games.  What happens when this kind of result occurs?

• One could simply say that the winning player was lucky enough to execute a swindle.  This may be true, but doesn't help our understanding of why and how this "luck" was created.
• Tenacity is essential for the initially losing and eventually winning player; for that reason, it was highlighted as its own category in my Chess Performance Inventory.  Annotated Game #104 was a great personal example, as I nearly gave up and resigned after Black's kingside breakthrough (which gave him an objectively won game), but made the deliberate decision to fight as hard as possible.
• Although refusing to give up the fight is a necessary condition, it's obviously not sufficient in itself to disrupt your opponent's play.  Creating threats by obtaining active play for your pieces is absolutely essential, otherwise your opponent will have no trouble finishing you off.  In other words, give them an opportunity to go wrong with their subsequent play by ignoring your threats (the same way you got in trouble in the first place, no doubt.)

On a more sophisticated level one can use what GM Alex Yermolinsky in The Road to Chess Improvement calls "trend-breaking tools".  For example, a pawn sacrifice in return for enhanced piece play can let you seize the initiative and get more practical chances, even if a computer engine might indicate no difference in the evaluation (or perhaps a slightly worse one).  One characteristic of master games is how highly valued the initiative is, as the side which possesses it is the one making threats and dictating the course of the game.  Psychologically speaking it is almost always harder to defend than attack, even when an objective evaluation of the position might be completely neutral.

Something that is often mentioned by chess trainers, especially for players at the Class level, is how aggressive players with less knowledge of the game tend to win much more often than players with greater skills and understanding, but who play passively.  Naturally a too-aggressive player can and will over-press their opponent and lose, but it's worth reflecting on how their performance over time will still overshadow that of a too-passive player, who simply does not create enough opportunities for themselves.  It's much easier for a winning, aggressive player to then fill in the skills and knowledge gaps and thereby improve, than it is for the passive player to break out of their rut.

My own historical "playing style" was certainly more on the passive side.  One of the key aspects of improvement for me has been to break out of that stereotype and my more recent games have reflected that - although it's obvious a good deal of work remains to be done on my skills.

Annotated Game #105: The Caro-Kann - a dull, positional opening (?)

This final-round Slow Chess League tournament game featured an imbalanced Caro-Kann, where Black staked everything on an audacious kingside attack after the dubious decision to sacrifice two pawns coming out of the opening.  Essentially I overreacted to White's 10. Qb3, an excellent move which nonetheless could have been neutralized with a little more thought.  However, it being a Friday night after a long week of work and a glass or so of wine, I had something of a devil-may-care attitude and felt like playing actively, rather than defending a potentially worse position with objectively best moves.

White, perhaps seeking to consolidate his advantage rather than press it, soon afterwards missed the powerful 16. Ng5! which while not winning immediately would give him an overwhelming advantage.  I was then able to continue with my attacking plans, although missing the much stronger 21...Bxh2+ continuation.  (Thanks to nate23 for raising this in the post-game analysis chat.)  White's king is eventually chased across the board and material loss ensues.

Black's ability to fight for the initiative and then put together a winning attack helps illustrate the practical power of active, aggressive play.  I constantly searched for threats to make against my opponent and for ways to improve my piece positions for attacking purposes, which increased the pressure on White.  Annotated Game #104 saw a broadly similar game trajectory, in which I was able to pull out a victory after being under pressure and down material.  While the point of post-game analysis is to help avoid putting oneself in that kind of a hole during future games, it's also worthwhile to understand on a broader level what kind of play can be most effective in terms of results.  I'll share some related thoughts on chess performance vs. chess skills in a subsequent post.

[Event "Slow Swiss #8"] [Site "Chess.com"] [Date "2013.09.20"] [Round "5"] [White "TheIronFly"] [Black "ChessAdmin_01"] [Result "0-1"] [ECO "B13"] [WhiteElo "1421"] [BlackElo "1550"] [Annotator "ChessAdmin/Houdini"] [PlyCount "64"] [EventDate "2013.??.??"] [TimeControl "45"] 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nf3 $5 (5. Nc3 {would enter the Panov variation.}) 5... Bg4 {I thought for a while here and decided to be experimental, rather than almost certainly enter the main Panov line with ... e6, which would transpose back after 6. Nc3. To me this was the best way to challenge the difference in move-order. Its drawback is that the bishop is away from the queenside; the effects will be seen in several moves.} 6. Nc3 e6 (6... Nc6 {would be better here for several reasons, developing the piece, freeing the Ra8 and and blocking the a4-e8 diagonal.}) 7. cxd5 (7. Qb3 { now looks pretty strong.}) 7... Nxd5 8. Be2 Be7 (8... Bb4 $5) 9. O-O O-O { played with almost no thought. I was still on automatic pilot in the opening, which is dangerous when the opening has varied from the standard.} (9... a6 { would have been the easiest way to avoid annoying threats to the b-pawn.}) 10. Qb3 {the most challenging move (and annoying for Black). I thought a good deal here and came up with several plausible options, in fact Houdini's top 3 choices (.. .Nc6, ...Nxc3 and ...Nb6) . I ended up not seeing any of their results clearly enough, however, and played an inferior move.} Nd7 $2 {Black is not going to have enough compensation for the sacrificed material.} (10... Nxc3 11. bxc3 Qc7 $11 {seems simplest. During the game I over-valued the idea of keeping the d-pawn isolated in my calculations.}) 11. Nxd5 exd5 12. Qxd5 Be6 $6 {rather than play a game where I am down a pawn and struggling the whole time, I decide to sacrifice another pawn for activity and a possible kingside attack. Did I mention I had been drinking wine prior to the game?} (12... Qb6 { would be a solid move that limits the damage.}) 13. Qxb7 Nb6 14. Qe4 $16 Re8 15. Be3 Bd6 $2 {Black abandons the defense of the diagonal in hopes of increasing his attacking potential, but this gave White a missed opportunity.} (15... Bd5) 16. Rac1 $2 {this gave Black a pawn back for free, essentially.} ( 16. Ng5 {this was pointed out in the post-game analysis by nate23.} f5 17. Qf3 Qd7 18. Nxe6 Qxe6 19. Rac1 {is one possible continuation. White's two bishops, extra two pawns and the passed central pawn, along with Black's weakened kingside, give White an overwhelming advantage.}) 16... Bxa2 $14 17. Qd3 Bd5 18. Ne5 {I thought this was a bit of a time-waster, although it does provoke a weakness on the g8-a2 diagonal which bears watching.} f6 19. Nc6 {the knight is further from the coming kingside action here, as well as being exposed to threats against it, so I was happy to see this.} Qd7 20. Na5 Re4 {perhaps not objectively best, but it seemed a good way to disrupt things and attempt to wrest the initiative away from White. The rook of course is headed for h4.} 21. Bf3 $2 {this looks superficially good but allows Black to execute the attack, even though I missed the best continuation.} (21. h3 Rh4 22. Bd2 Re8 $14) 21... Rh4 $15 (21... Bxh2+ $1 {was also pointed out by nate23 in the post-game chat. On most other moves I looked at the bishop sacrifice possibilities, but did not here, because I was fixated on the ...Rh4 idea. The general idea behind the rook move is correct, but the intermediate sacrifice gives it much more strength.} 22. Kxh2 $2 (22. Kh1 $19) 22... Rh4+ 23. Kg1 Bxf3 {and White cannot avoid mate, since the Black queen cannot be kept out of h3.}) 22. Bxd5+ Nxd5 23. f4 {this seemed like a mistake at the time, but Houdini assesses other moves as worse. White simply has no good choices.} (23. h3 Rxh3 24. gxh3 Qxh3 $17) 23... Nxf4 {I wondered here if the bishop capture would be better (Houdini says it is) but after looking at the knight capture I decided it looked fine and I did not want to take the time to work through the other variation. Lazy, or practical?} (23... Bxf4 24. Bxf4 Nxf4 25. Qb3+ {forcing Black to have to worry about his back rank} Kh8 26. Qf3 Re8 $15) 24. Bxf4 $6 ( 24. Qc4+ {I had considered this to be White's most effect move here.} Kh8 25. g3 Nh3+ 26. Kg2 Qg4 $11) 24... Bxf4 25. Rc4 $2 {I was expecting g3 here. The move is a creative attempt to hold the position together, but after Black's next move, White cannot avoid losing material while his king gets chased to the other side of the board.} (25. g3 Bxc1 26. gxh4 Bh6 27. d5 Rb8 $15 { is one probable continuation.}) 25... Bxh2+ 26. Kf2 Rf4+ 27. Ke2 Re8+ 28. Kd2 Rxf1 {this seemed the most sure way to victory, although there were better options.} (28... Qd5 $5 {forks the Na5 and g2, for example.}) 29. Qxf1 Qb5 30. b4 Bd6 {I had calculated this far before move 28 and did not see any way for White to hold out.} 31. Nc6 Bf4+ {a useful example of how my thinking process was working well by this point in the game. This non-obvious move works because of the deflection tactic. If the king moves, then Black's rook and queen penetrate with mate threats.} 32. Qxf4 Qxc4 0-1

Annotated Game #104: Queenside? check. Center? check. Kingside? fail

This round 4 game from a Slow Chess League tournament shows the swift punishment that can occur when the kingside is neglected, regardless of what else is happening on the board.  The English Four Knights opening allowed me (White) to establish strong central pressure and at the same time restrain Black effectively on the queenside.  However, my plan to establish a battery against Black's king to increase the pressure was too slow and an ill-advised queen maneuver allowed Black to quickly marshal what looked like a crushing attack.

Despite feeling a certain amount of despair at my predicament, I was determined to actively defend and not allow Black to achieve an easily winning position.  After a harrowing sequence, I emerged down two pawns but having avoided the mate threats.  Continuing to play as actively as possible and make threats with my queen in the initial queen and minor piece endgame, I was able to catch Black in a tactic and suddenly I had the only piece left on the board, a mighty dark-square bishop, which combined with my king was enough to sweep up all the Black pawns and force the queening of the last White one.

While the outcome was eventually positive, next time a bit of prophylaxis on the kingside would not hurt...

[Event "Slow Swiss #8"] [Site "Chess.com"] [Date "2013.09.12"] [Round "4"] [White "ChessAdmin_01"] [Black "Yamaduta"] [Result "1-0"] [ECO "A28"] [WhiteElo "1504"] [BlackElo "1504"] [Annotator "ChessAdmin/Houdini"] [PlyCount "137"] [EventDate "2013.??.??"] [TimeControl "45"] {A28: English Opening: Four Knights Variation} 1. c4 e5 2. Nc3 Nf6 3. Nf3 Nc6 4. e3 Be7 5. d4 exd4 6. Nxd4 Ne5 {this unusual move seemed to me to be most likely a prelude to ...c5, to kick the Nd4, but it never actually occurred in the game. The Ne5 looks nicely centralized, but it doesn't seem to do enough to warrant the tempo loss of moving the piece twice.} (6... Bb4 7. Nxc6 Bxc3+ 8. bxc3 dxc6 9. Qc2 Qe7 10. f3 Be6 11. Rb1 b6 12. Qa4 Bd7 13. Be2 c5 14. Qc2 O-O 15. O-O Bc6 16. e4 Rad8 17. Bg5 h6 18. Bh4 g5 19. Bf2 {1/2-1/2 (19) Gulko, B (2560)-Ivanov,A (2555) St Martin 1992}) (6... O-O {followed by ...d5 is the main line and was used most famously by Karpov.}) 7. Be2 {White continues with standard development, as occurs in the main line.} (7. Nc2 {was played in the only master-level game in the database. The idea is to regroup White's pieces to exploit an eventual f4 advance.} d6 8. e4 O-O 9. Be2 c6 10. O-O Re8 11. Kh1 Qc7 12. f4 Ned7 13. Bf3 Bf8 14. Re1 a5 15. b3 Nc5 16. Ba3 Ne6 17. Qd2 Nd7 18. Rad1 Nec5 19. g3 Rb8 20. Kg2 Ra8 21. h4 Rb8 {Suba,M (2487)-Voos,G Mallorca 2000 1-0 (36)}) 7... O-O 8. O-O d6 9. b3 {fortifying c4 and getting the bishop to the long diagonal.} b6 $6 {this weakens the light square complex on the queenside and I was not too concerned about the bishop on the long diagonal, believing that I could exchange it off via an eventual Bf3. Houdini points out the tactical exploitation I could have made of the light square weakness; unfortunately, I was not even considering the possibility of an f4 push (which otherwise would be antipositional). The vulnerability of the Ne5 combined with the light square weakness is what makes the tactic work.} 10. Bb2 (10. f4 Ng6 11. Bf3 {and Black loses material.}) 10... Bb7 {White can proceed in a number of ways here. I chose to retreat the knight to block the long diagonal and then develop with Qc2.} 11. Nf3 Ng6 12. Qc2 {with this move, Black no longer controls e4 and White can connect the rooks.} d5 $6 {this gives White a big target and pressure on the d-file, I assume my opponent missed the pin idea and assumed the pawn would be immediately exchanged.} 13. Rfd1 c6 14. e4 { I thought for some time here and figured that increasing the pressure on d5 was best. Houdini is less impressed with the idea, preferring to use the tempo to begin shifting pieces to more effective squares. A problem with the text move is that it further boxes in the light-squared bishop.} (14. Bd3 $5 { this at first seems to obscure the pin, but in reality it makes it more dangerous, as the threat is Bxg6 with a discovered attack on the d-file, for example after .. .dxc4.} Re8 15. Rac1 Bd6 16. cxd5 cxd5 17. Bf5 $14 {is one possible continuation, leaving Black with the isolated queen's pawn.}) 14... Rc8 {I had thought that Black would do best to immediately move his queen away from the pin, but evidently he wanted to develop his rook and then have potential counterthreats on the c-file.} 15. Rd2 {not a very creative approach, albeit consistent, aiming to double rooks and further increase the pressure.} ( 15. Bd3 dxc4 16. bxc4 Qc7 17. e5 $14 {is an interesting idea from Houdini. I had given some thought during the game about the idea of the e5 thrust, but didn't find a way to make it meaningful. Here the Q+B battery adds some significance to the move.} Nd7 18. Bf5 {here the bishop is much more powerful than in the game continuation.}) 15... Qc7 16. Rad1 Rfd8 17. Qb1 $6 {the idea was to enable the exchanges on d5 without exposing the queen to tactics along the c-file. However, this was unnecessary and it potentially gives away the initiative to Black.} (17. e5 {here also, the idea of the pawn thrust could be useful.} Ng4 18. cxd5 cxd5 19. Nxd5 Qxc2 20. Nxe7+ Kf8 21. Rxc2 Rxc2 22. Rxd8+ Kxe7 23. Rd2 $14) (17. exd5 cxd5 18. cxd5 Bb4 19. Rd4 (19. Bc4 {would also be OK, with the idea of disrupting the c-file pressure.}) 19... Bxc3 20. Rc4) 17... Bc5 (17... Bb4 {I thought would have been more to the point, exploiting the inaccurate queen move.} 18. exd5 (18. a3 $2 {leads to problems for White.} Bxc3 19. Bxc3 Nxe4 20. Rc2 dxc4 21. Bxc4 c5 $17 {and now Black is a pawn up with the Bb7 a real monster.}) 18... cxd5 19. Qc2 {and White has accomplished little, other than handing the initiative to Black.}) 18. exd5 cxd5 19. cxd5 a6 {preventing Nb5, which is what I had in mind.} 20. a4 {employing some prophylaxis and taking control of the b5 square, as I was concerned about Black playing ... b5-b4 otherwise.} Qf4 {this seemed rather cheeky at the time (and objectively is so), but I react poorly and manage to ignore all of Black's threats in favor of following my own short-term plan of creating a Q+B battery.} (20... Bb4 $5) 21. Qa1 $2 (21. Bc4 {what a difference a tempo makes, as this played immediately would give White a fine game by shoring up d5, blocking tactics down the c-file against the Nc3, and laterally protecting f2 via the Rd2. The tempo wasted by the text move nearly sinks me.}) (21. h3 { is what I thought would have been the simplest way to prevent Black's next sequence, but in fact Black has other tactical resources. In practical terms, though, it would be difficult for someone to see the following sequence.} Bxf2+ 22. Kxf2 Rxc3 23. Bxc3 Ne4+ 24. Kg1 Nxc3 25. Qc2 Nxd1 26. Qxd1 Qd6 $15) 21... Ng4 22. Bc4 $2 {now this move should lose. As soon as I played it, I saw Black's next coming and despair started to set in.} (22. Rd4 {to my credit, I had in fact considered this exchange sacrifice as a defensive resource. Had I seen Black's 22nd move in the game, I probably would have played it.} Bxd4 23. Rxd4 Qf5 24. h3 Nf6 25. Bd3 Qd7 26. Qd1 {and White is OK.}) 22... Nh4 $1 $19 23. Ne4 {I thought for some time here and this seemed the best practical chance for the defense. Houdini agrees (although it's still losing).} Nxf3+ 24. gxf3 Qxh2+ 25. Kf1 Qh3+ {now Black fails to bring more pieces into the attack, which is what allows White to escape.} (25... Re8) 26. Kg1 Qxf3 27. Nxc5 Bxd5 $2 {this gives White an out by allowing him to eliminate the bishop. I took it too quickly in the game, however, not considering the intermediate move Houdini points out.} (27... Rxc5 $19 {and now Black can sacrifice on d5 to eventually deflect the Rd2 (a necessary defender of f2), which wins.} 28. Be5 Rcxd5 29. Bxd5 Rxd5 30. Rxd5 Qxf2+ $19) 28. Bxd5 (28. Ne4 Qxe4 29. Rxd5 Rxd5 30. Bxd5 Qf4 31. Bd4 {and the best Black can do is a perpetual check.} Qh2+ 32. Kf1 Rc2 33. Bg2 Qf4 34. Kg1 Qh2+) 28... Rxd5 29. Ne4 {a recurring active more for the defense.} Rxd2 30. Rxd2 {White remains tied to the defense of the f2 pawn.} Qxe4 31. Qd1 {White is still losing, down two pawns and with an open king position, but is no longer in imminent danger of being mated. Black now has to think about his own king position due to the back rank.} Re8 (31... h5 { immediately is what Houdini recommends, which would rid Black of his bank-rank weakness and see him start to leverage his kingside pawn majority.}) 32. Rd8 h5 33. Rxe8+ Qxe8 34. Qd4 {White's strategy here is to play as actively as possible with his queen and continue making threats, to try and keep Black from consolidating his advantage.} Qe1+ $6 (34... f6 {permanently blocks the mate threat.} 35. Qd5+ (35. Qxb6 $2 Qe1+ 36. Kg2 Qe4+ 37. Kg1 Qb1+ 38. Kg2 Qxb2 ) 35... Kh8 $19) 35. Kg2 $15 {White's king still looks precarious, but Black can do nothing further against it and now must deal with White's threats.} Nf6 36. Qd8+ Qe8 $2 {Black misses the next tactic, which turns the tables and gives White a winning game.} 37. Bxf6 Qxd8 38. Bxd8 $18 {in the bishop ending, White's strategy is to restrain the Black pawn majority, then penetrate and destroy it with the king, using the bishop for the necessary extra tempi to put Black in zugzwang.} b5 39. Kg3 g6 40. axb5 axb5 41. Kf4 Kg7 42. Kg5 { the Bd8 dominates the d8-h4 diagonal and effectively shackles Black's position, preventing him from driving away the king.} f5 43. Bb6 Kf7 44. b4 {White is in no hurry and places the pawn on a dark square, so it can be defended in the future by the bishop if necessary.} Kg7 45. Bd4+ Kf7 46. Bc5 Kg7 47. Bb6 { the bishop is moved for a tempo gain, forcing Black to allow White's king to penetrate.} Kf7 48. Kh6 h4 49. Bc7 {ensuring that the h-pawn has no future.} h3 50. Kg5 Kg7 51. Kh4 Kf6 52. Kxh3 g5 53. Kg3 g4 54. f3 {this will allow the liquidation of all kingside pawns, making White's task easier.} Ke6 55. fxg4 fxg4 56. Kxg4 Kd5 57. Bd8 Kd6 58. Kf4 Kd5 59. Be7 {with the White pawn immune from attack, White can repeat the zugzwang trick again, as Black's king cannot protect his b-pawn forever.} Kc6 60. Ke4 Kb6 61. Kd5 Ka6 62. Kc6 Ka7 63. Kxb5 Kb7 64. Bh4 {with the bishop present to provide a tempo to allow White to gain the opposition, it is an easy win.} Kb8 65. Kb6 Ka8 {playing for a stalemate chance.} 66. b5 Kb8 67. Bg5 Ka8 68. Kc7 Ka7 69. b6+ 1-0