Canadian Chess Master

Robert Sasata (2350) - IM Torsten Bae (2383)
Battle at the Border (4)
Lloydminster 2007
D70

1. d4 Nf6 2. c4 g6 3. f3 d5 4. cxd5 Nxd5 5. e4 Nb6 6.Nc3 Bg7 7. Be3 O-O 8. Qd2 Nc6 9. O-O-O

Does White not have the ideal attacking formation against the king’s fianchetto? It feels like the position plays itself…

9…  f5

As recommended by Rowson in his Understanding  the Gruenfeld. Personally, I think Black is struggling to maintain equality now.

10. h4 e5 11. d5 Nd4 12. h5

I was secretly hoping for 12…fxe4 13.fxe4 Bg4 with a transposition into my Haessel game. Bae increases the complexity and tosses another
pawn lever into the mix.

12… c6 13. hxg6 hxg6 14. exf5

Psychological draw offered by me here. I was pretty sure he wouldn’t accept it, at such a complex position at an early stage of the game. So why offer? Sometimes the draw offer has interesting secondary effects, such as:
- making the draw refuser bear the conscious or subconcious burden of having to justify his non-acceptance by demonstrating their position to be ‘more than equal’
- making the opponent use valuable thinking time contemplating possible tournament placings, prize money distributions, etc.
- giving the opponent the false impression of ‘fear’ on my part. Offering the draw may make him think that I don’t really like my position that much (a usual reason for offering a draw) and perhaps cause him to overestimate his chances.
Or, none of the above. Obviously, this ‘parlour trick’ should have little effect against a seasoned veteran, but one must not underestimate the cumulative effect of subtle psychological ploys such as these.    Back to the board and pieces. The preponderance of captures make this position very complex and difficult to evaluate without a deep and extensive analysis, the time for which I don’t have, so I’ll just resort to the standby of giving the engines’ suggestions and evaluations:

14. dxc6 bxc6 15. exf5 Bxf5 16. g4 Bc2 17. Bxd4 exd4 18.Kxc2 dxc3 19. Qxd8 Raxd8 20. Rxd8 Rxd8 21. bxc3 +/= {0.23/15}
14. Bg5 Qd6 15. Bh6 f4 16. dxc6 bxc6 17. Qf2 Bxh6 18. Rxh6 Qe7 19. Bd3 Kg7 20. Rh1 Be6 {0.02/15}
14. Nge2 Nc4 15. Qd3 Nxe3 16. Qxe3 c5 17. exf5 gxf5 18. f4 e4 19. Qg3 Qb6 20. Kb1 Bd7 {0.01/15}
14. f4 fxe4 15. dxc6 Qe7 16. Nxe4 exf4 17. Bxd4 Qxe4 18.Bxg7 Kxg7 19. cxb7 Bxb7 20. Qc3+ Rf6 {0.00/14}
14. Bh6 f4 15. Bxg7 Kxg7 16. Nge2 c5 17. Nxd4 cxd4 18. Nb5 Bd7 19. Kb1 Bxb5 20.Bxb5 Qd6 {0.27/15}) 14… cxd5

14… cxd5

More number-crunched variations for the fanatics:

14… Bxf5 15. dxc6 bxc6 16. g4 Bc2 17. Bxd4 exd4 18. Kxc2 dxc3 19. Qxd8 Raxd8 20. Rxd8 Rxd8 21.bxc3 {0.26/14}
14… cxd5 15. fxg6 Bf5 16. Bd3 Nc4 17. Bxc4 dxc4 18. Nge2 Bd3 19. Bxd4 exd4 20. Ne4 Qe8 21. Rh4 {0.35/14}
14… Nxd5 15. Nxd5 cxd5 16. fxg6 Bf5 17. Bd3 Qf6 18.Bxf5 Qxf5 19. Qd3 Qxd3 20. Rxd3 Rac8+ 21. Kb1 {0.38/14}
14… Rxf5 15. dxc6 Qc7 16. Bd3 Rf6 17. cxb7 Bxb7 18.Nge2 Rd6 19. Bh6 Bf6 20. Qe1 Nf5 21. Kb1 {0.50/14}
14… Qf6 15. g4 gxf5 16. g5 Qg6 17. f4 c5 18. Qg2 exf4 19. Bxf4 Bd7 20. Nf3 Rfe8 {0.63/14}
14… Nxf5 15. Bg5 Qd6 16. dxc6 Qxd2+ 17. Kxd2 Ng3 18. Rh4 Nxf1+ 19. Rxf1 bxc6 20. Ne4 Nd5 21. Ne2 {0.15/15}

15. g4!

It takes over 10 minutes for Rybka to agee that this is White’s best move. It took me far less :)  The second rank gets opened for Qh2.

15… gxf5!

Anything else is inferior.

16. Qh2 Qf6?!

Bae responded with this move rather quickly. Interestingly, this move is also Rybka’ first choice, but given time to get to 15 ply, it drops to fourth place, with a 0.68 evaluation. Offered as best play is 16… Qe7 ! 17. Qh7+ Kf7 18. Qh5+ Kg8 (only move) 19. Re1 Nc6 (only move) 20. Qh7+ Kf7 21. gxf5 (21. Rh6 Ke8! - Black deftly sidesteps White’s kingside forces, and takes temporary shelter behind the central preponderance.
(21… Rh8 is not as good, as depicted in this tortuously complex, variation: 22. Qg6+ Kf8 23. Rxh8+ Bxh8 24. Nh3 Qf7 (24… d4 25. Ng5! Qe8 26. Nh7+ Ke7 27. Bg5+ Kd7 28. Nf6+ Bxf6 29. Qxf6 a6! {Preventing a deadly Nb5+ in various lines.} 30. gxf5 Kc7 31. Ne4 +/-) 25. Bc5+ Ke8 26. Qd6 Qe7 27. Nb5 Kf7 28. Bd3 Qxd6 29. Nxd6+ Kf6 30. gxf5 +/-)
22. Rg6 Bf6 23. Qh5 Kd8 with balanced chances in a highly complex position.) 21…Qf6 (only move) 22. Qh5+ Kg8 23. Bxb6 axb6 24. Nxd5 Qxf5 =. Just in case there’s any confusion, I should explain that I did not see any of this stuff at the board. I was mostly playing by intuition, trying to keep my tempo brisk, for practical purposes.

17. Qh7+ ?!

Since this forcing move peters out to eventual equality, it must be adorned with a ?! diacritical. Given sufficient time, Rybka demonstrates that the sophisticated knight-undermining 17.f4! keeps an advantage, depicted here in this long complex variation difficult for normal humans to find: 17. f4! fxg4 18. Qh7+ Kf7 19. fxe5 Qxe5 20. Bxd4 Qg5+ 21. Rd2 Rh8 22. Qxg7+! Qxg7 23. Rxh8 Qg5 24. Rh7+ (+/-) Kf8 25. Bd3 Bf5 26. Rg7 Qf4 27. Nge2 Qf3 28. Bxf5 Qxf5 29. Ng3 Qf3 30. Rg5 Qxg3 31. Rf2+ Ke8 32. Rg8+ Ke7 33. Rg7+ Ke8 34. Bc5 +-.

17… Kf7 18. Rh6! Rh8 (only move)

If 18… Qe7? 19. Qg6+ Kg8 20. Bg5 Qc7 21. Qh7+ Kf7 22. Rg6 wins.

19. Rxf6+ Kxf6 20. Qxh8

The alternative 20. g5+ Kf7 21. g6+ Kf8 22. f4 Rxh7 23. gxh7 Nc6 is just equal.

20… Bxh8 21. Rxd4

White can try to recycle the f4 idea, but it also leads to equality:  21.f4 Kf7 22. fxe5 Nc6 23. gxf5 d4 24. Bxd4 Nxd4 25. Rxd4 Bxe5 26. Re4 Kf6 27. Nf3 Bxc3 28. bxc3 Bxf5 29. Rd4=.

21… Be6!

When I first played Rxd4, I thought I had the advantage and anticipated something like 21… exd4 22. Bxd4+ Kg6 23. Bxh8 fxg4 24. Bd4 with good chances for the full point after a long endgame grind. As I sat there thinking, it dawned on me that …Be6 was a very good move, and that the game complexion would shift towards equality. It cost him about half an hour on the clock, but he found it.

22. Rd1

Weaker is 22. g5+ - White must be careful to not press to ambitiously - Kf7 23. Rh4 d4 =/+.  It is important that the strength of Black’s pawn center be dissolved by returning the piece.

22… d4 23. gxf5 Bxf5 24. Bxd4 exd4 25. Rxd4 Ke7

Black two bishops and better coordinated pieces perfectly compensate for the pawn deficit. I was happy to agree to Bae’s draw offer.
1/2-1/2

During the Christmas holidays last year I traveled to Edmonton to play in the "Weekend Before Xmas" Team tournament, along with team mates Jeff and Jamin. The event was well-attended and an enjoyable experience. Staying with friends in Edmonton meant that little sleep was to be had during the evenings, but remarkably my chess didn’t seem to suffer too much.

The following game was a critical one in this tournament. Our team, the "Benko Brothers" faced the other leaders, the "Raging Kramniks", with whom we were tied.

Yearwood (2154) - Sasata (2321)
Weekend Before Xmas Team Tournament (4)
Edmonton, 2006

1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Bxc6 dxc6 5. O-O Bg4 6. d3 Qf6 7. Be3 Bxf3 8. Qxf3 Qxf3 9. gxf3 Bd6 10. Nd2 Ne7 11. Nc4 Ng6 12. Kh1 O-O 13. Rg1 Be7 14. d4 exd4 15. Bxd4 Rad8 16. Bc3 f5 17. Ne5 fxe4 18. Nxg6 hxg6 19. Rxg6 Bf6 20. Rag1 Bxc3 21.bxc3 Rd7 22. fxe4 Rxf2 23. R6g2 Rf4 24. Rg4 Rxg4 25. Rxg4 Kf7 26. Rg3 Ke6 27. Kg2 Rd2+ 28. Kh3 Kf6 29. Rg2 Rd7 30. Kg3 g5 31. Kf3 Rh7 32. Rf2 Rh4 33. Ke3+ Ke6 34. Rg2 Rh5 35. Kd4

Let’s join the action here, after White’s 35th move. Superficially, White’s position looks disastrous, with weak pawns all over the place. Closer examination reveals, however, that Black’s task is not that simple. White’s centralized king keeps his counterpart at bay, and White can defend his pawns in a tenuous equilibrium. During the game I thought that Black had no more than a small advantage.  However, by this time my teammates’ games had both finished, and we were sitting at 0.5/2, so I had to try to make something happen.

To make progress, at some point Black has to try to win the pawn at a2, even if that means allowing White’s rook counterplay on the seventh rank. I decided to slowly advance the queenside pawns before undertaking these measures.

35… b6 36. Ke3 c5 37. Rg3

Interesting… Roy ‘taunts’ me to take on h2, and provides easy access to the rest of the weaklings. What is the method to his madness?

37… Rxh2 38. Rxg5 Kf6?!

At the time, the obvious follow-up capture seemed to allow White dangerous counterplay: 38… Rxc2 39. Rg6+ Kd7 40. Rg7+ Kc6 41. Rg6+ Kb7 42. Kd3 Rxa2 43. e5. However,  my buddy Rybka assures me that the danger was merely illusory.  For example, 43.. a5 44. e6 Kc6 45. e7+ Kd7 46. Re6 Ke8 47. Rc6 Kxe7 48. Rxc7+ Kd6 -/+

39.Rf5+

Okay, so one winnning opportunity has passed, but I still have some positional advantage, and carry on with the original game plan.  However, White should not have major difficulties holding the position.

39… Ke6 40. Kd3 Rh3+ 41. Kd2 a5 42. Rg5 Rh6 43. Kd3 c6 44. Ke3 a4 45. Kf4 Rh4+ 46. Rg4!?

Though I managed to maintain a calm outward demeanor, internally, I was floored this move. It seemed to me that White had just played a horrible blunder that lost instantly, and it seemed so incredibly obvious. All Black has to do is trade rooks, advance the queenside pawns, and then sacrifice with a …b3 push for a queening breakthrough. To stop the a-pawn, White’s king would have to give up protection of e4, resulting in a simple winning king and pawn ending. Easy, right? Wrong!

46… Rxg4+ 47. Kxg4 Ke5 48. Kf3 c4 49. Ke3??

Only now does White make the game-losing move, although I was oblivious at the time. Amazingly, if White stops my a-pawn by playing a3 himself, Black cannot win this ending! Witness: 49. a3 c5 50. Ke3 b5 51. Kd2 (forced - White must be in the ’square’ of the a-pawn before Black plays …b5-b4) 51… Kxe4 52. Kc1.  I invite you to confirm for yourself the drawn nature of this position. No amount of triangulation or move-losing will alter the fact that White can maintain the
opposition if I manoeuvre my king to the d1-d2 squares, so the only remaining try is something like 52… b4 53. cxb4 cxb4 54. axb4 c3 55. b5 Kd5 56. Kb1 Kc5 57.Ka1 Kxb5 58. Ka2 =.

49… a3 50. Kf3 c5 51. Ke3 b5 52. Kf3 b4 53. cxb4 cxb4 54. Ke3 b3 0-1

Just as I, uh, ‘planned’.  Roy, obviously displeased with the outcome, took out his frustrations on a defenseless pencil. A very instructive ending for me, and I hope you learned something from it as well.

A move characteristic of my ‘take it easy’ approach to the QGD variations back in the day. Lines where White plays Bg5 are more aggressive, and perhaps more testing of Black’s theoretical preparations, but I enjoy playing the restrained approached where the c1 bishop is left at home for a while. Transpositional possibilities abound here: 4…dc4 5.Bc4 c5 leads to a Queen’s Gambit accepted, 4…c6 is the Semi-Slav, 4…c5 is a Tarrasch, while even the plausible 4…b6 leads to (an inferior line of) the Queen’s Indian Defence. Black’s move further delays having to make a choice about the opening.

Inviting Black to ‘win’ a tempo and clarify the central pawn structure. But the loss of tempo is compensated by the fact that Black has already committed his knight to d7, and so must play an inferior version of the Queen’s Gambit accepted. Here, if Black plays 5…c6, hoping to transpose in the Meran after 6.Nc3, I can take advantage of my move order and play 6.Ndb2, so that I have the option of recapturing on c4 with the knight. White then ends up with a comfortable middlegame position after the eventual e3-e4 push.

By placing the bishop to d6 Black signals that he in aiming to develop a setup where, after the queen goes to e7, the e5 thrust will be enabled, and Black will be well-centralized. The plan makes even more sense considering that b5 square is covered by the a6 pawn, and so White will not be able to harass the d6 bishop with Nb5.

My play revolves around using my positional trump: the central pawn majority. The planned e3-e4 advance will also solve the problem of the inactivity of the c1 bishop.

More logical than the alternative, less successful ways approaches to using the central pawns:

11. Nxe5 Nxe5 12. dxe5 Bxe5 13. f4 Bxc3! (13… Bd6?! 14. e4+/-) 14. bxc3 O-O=

An excellent square for the knight, where it attacks the bishop and probes the sensitive g7 square. The mere presence of this looming terror causes Black’s position enough discomfort such that over the next few moves spent ejecting this piece, the equilibrium shifts and Black’s pieces lose coordination.

Black removes the d6 bishop from attack, and threatens to win the e4 pawn by removing the guard at c3.

Ignoring the threat, and increasing the pressure on Black’s position.

Black takes the booty. It’s generally considered worthwhile to make an extra effort to nab a central pawn, and in addition, this move also forces White to make a choice about the f5 knight.

Black retreats to the kingside, lending additional support to f7, and preventing Nh5. There is a drawback to this placement however…

Taking advantage of the bishop’s lack of squares. Black must now tread very carefully in this position to avoid drifting into inferiority.

I don’t like this move; Black is not gaining a tempo by attacking the queen, rather, he encourages White go to the square he wanted anyways, and at the same time removes support for the potentially vulnerable f7 square.

My chess engines suggest that 17… h6 18. Bh4 Re8 19. Qa2 Qe7 20. f5 Bh7 21. Nh5 Ne5 is a better defence, but I still prefer White.

Black tries to initiate counterplay and distract White from the pressure on the diagonal leading to f7.

After this move Black’s position falls apart. The best continuation was

20… Bd3! with the idea of retreating the bishop to b5 and alleviating the pressure on the a4-e8 diagonal. 21. Bxf6 Qxf6 22. Bxd7 Red8 23. Ba4 Bxf1 24. Rxf1 Qxc3 25. Bb3 Rd7 with equal play.

White forces a queen trade and exploits the vulnerability of c6.

22… Qxf2+ 23. Rxf2 Bh5 24. Bxc6 Ra7 25. Bxf6 1-0 [Robert Sasata]

I had just taken up the Dragon a couple of months before this game was played, and was eager to show off my many moves of memorized theory in the main lines of the Yugoslav attack. Here, Hjartarson quelled my youthful ambitions with the solid

A good practical choice for the first round of an open swiss; White avoids heavy theoretical variations against a young unknown (me) and makes the game more positional in nature. As it turned out, I was only vaguely familiar with the line, exemplified later by my planless middlegame play.

Popularized in the 1980’s by Karpov as an alternative to 8.Be3, this moves sucks dry much of Black’s active play and intends to develop a grip on the position.

A standard move, preventing f5, and planning …Rd8 with ideas of …d5. However, I remember not really feeling comfortable in this position due to inexperience.

13… Nd7 14. Kh1 Re8 15. Rd2 Nb6 16. Nd5 Bxd5 17. exd5 Nb8 18. c3 h6?!

Today I would have avoided this unnecessary voluntary weakening move. Years of experience accumulated in the intervening years has shown me that little pawn moves like this often have serious consequences in the future. Better was 18… N8d7+/=

Black’s game falls apart completely after this, although the position was difficult anyways. Now Johann hurts me.

I was in the lead with 2.5/3 going into this round. Before the game started, I learned that my last round opponent had withdrawn from the tournament that morning, effectively giving me (and my nearest contender, Jack Yoos, who had 2/3) a full-point last round bye. I tried not to let this knowledge affect my play, but I did contemplate offering a draw (around the time I made my speculative 17th move) to ensure a share of first, but I changed my mind and instead played the move silently. In retrospect, I believe this decision to have been fortuitous, as a draw offer might well have been perceived as psychological weakness, and may have inspired Gardner to look harder for the variation that I was afraid of. As it went, I think the (naive) outward confidence with which I played the move, combined with Gardner’s looming time trouble, was enough to swing the psychological balance in my favour. I was expecting Gardner to play the Dutch Defence, and had decided to try an ‘offbeat’ but dangerous attacking line to pose concrete problems right out of the opening. The Leningrad Dutch Defence was in my own Black repertoire for several years, and I remember the lines I often felt uncomfortable playing against were the ‘irregular’ ones, such as

1.d4 f5 2.g4!?
1.d4 f5 2.h3!?
1.d4 f5 2.e4
1.d4 f5 2.Bg5
1.d4 f5 2.c4 g6 3.h4!?

and others of that ilk. I think the opening choice worked well, as Gardner spent a lot of time in the opening, and his subsequent time pressure was probably an instrumental factor in enabling my faulty combination to succeed.

1. d4 f5 2. Nc3 d5 3. e4!?

The move that Black’s second was intended to prevent, and the perfect line to get my opponent out of his comfort zone.

3… dxe4

Several years ago I caught an opponent unprepared for this line, and he responded by quickly centralizing his knight as follows: 3… Nf6 4. e5 Ne4 5. Nce2 I was pleased with this move at the time, and I thought it was a clever way to trap the e4 knight, who suddenly lacks escapes squares. The game continued 5… e6 threatening to escape via g5 6. h4 Be7 7. g3 c5 8. f3. Here Black does in a fact have some compensation the piece because the time spent and the structural concessions that I’ve made to win the knight have reduced my piece coordination. But, my opponent, probably psychologically demoralized for having lost apiece so quickly and unexpectedly, failed to put up much resistance. If Black just continues with 8… Nc6 9. fxe4 fxe4 10. c3 O-O 11. Nh3 the position is better for White, but there remain problems to be solved. 3… fxe4?! 4. Qh5+ g6 5. Qxd5+/- is a position I had in an email game.

4. Bf4

The immediate 4. f3 is considered less accurate, because after the central counter 4… e5 Black has at least equality. 4… e6 Black prepares to challenge the bishop’s active post with …Bd6. I wasn’t overly impressed with this move, and the subsequent positions Black has to put up with. In this position Black has tried stuff like 4… Nf6 5.f3 and allowed White to trade pawns on e4. The pawn structure is messy, but at least Black gets his pieces out more easily.

5. f3 exf3?!

I think this exchange makes things easier for White. I think he should play 5… Nf6 and after 6. fxe4 fxe4 the pawn on e4 hampers White’s development a little by controlling the f3 and d3 squares, and Black can think about moves like …Bb4 or …Bd6 and …Nc6 with fully equal chances in an interesting position.(But not the knight capture: 6… Nxe4? 7. Nxe4 fxe4 8. Qh5+ g6 9. Qe5 Rg8 10. Qxe4+/-)

6. Nxf3

Now White has easy development, and Black has a permanent positional weakness - a backwards pawn on a semi-open file.

6… Bd6 7. Ne5

The knight occupies the natural outpost on e5. Of course, I would be happy for Black to trade it off for his dark-squared bishop and be left with only the remaining bad bishop at c8.

7… Nf6 8. Bc4 c6

Black shores up his defence of d5 with the intent of placing the knight there to neutralize the c4 bishop and its uncomfortable pressure on e6. It also allows for the possibility of the queen developing to b6.

9. O-O O-O 10. Qd3

I spent a little time weighing the relative merits of various other queen placements, such as d2, f3, e2, or e1, in addition to prophylatic moves like Kh1, a4 or Bb3. How to choose between these moves? White’s lead in development suggests that it is ideal to bring the last piece into play (i.e. via Rae1) and thus the three prophylactic choices are too slow (and unnecessary - why prevent the structurally weakening …b5?). Additionally, I felt it useful that d4 pawn remain protected to reduce annoyance from a potential …Qb6. Of the two remaining queen moves, Qd3 also protects the central square e4, and threatens, in passing, Qxf5.

10… Nd5 11. Rae1!

My opponent revealed after the game that he hadn’t expected this move. Optically, the move is obvious, as it is the logical conclusion to the developmental scheme outlined in the last note. However, it takes some thought to part with the bishop on f4, as this trade implies the loss of the two bishops, and potential loss of dark square control. In addition, this bishop prevents the developing …Nbd7 (due to Nxc6). But look at the payoff: the c4 bishop gains in strength, the pressure on the e6 pawn weakness is renewed, and Black trades his only well-placed piece!

11… b5 12. Bxd5?!

Honestly, my candidate move list here consisted of only 12.Bb3 and the text. I didn’t realize that 12. Nxd5! was playable, missing the idea that after 12… bxc4 (12… exd5 13. Bb3) 13. Nxc4 attacks the d6-bishop twice, thus avoiding losing a piece, and maintaining a superior position. After the text White is still better, but the magnitude of the advantage is diminished. 13… cxd5 14. Nxd6 Ba6 15. Qe3+/-

12… exd5 13. Ne2!

Manoeuvring the knight to the far more useful square on d3.

13… a5 14. Qf3

Squares g3 and h3 were also locations I considered, but on f3 the queen attacks d5 and sets up tactical possibilities, such as Nxc6.

14… Qc7 15. Nc1 b4

Black gains some space on the queenside and prepares to activate his bishop via a6.

16. Ncd3 Ba6 17. Nxc6?!

In the game, this combination worked beautifully. However, it is based on an oversight! I should have played something more patient like Qh3 or Qh5 intending a rook lift with Rf1-f3.

17… Qxc6

Not 17… Nxc6? 18. Qxd5+ Qf7 19. Qxd6+/-

18. Re6 Bxd3?

18… Qxc2! I overlooked this rather obvious move in my original analysis of 17 Nxc6. In retrospect, this oversight was probably a good thing, as I could easily have burned up the majority of my remaining time trying to calculate the consequences . I think with best play by both sides the play is complex, but equal: 19. Rxd6 (19. Bxd6? Qxd3 20. Qxd3 Bxd3 21. Rd1 Be4 22. Bxf8 Kxf8 23. Rc1 Nd7 24. Rc7 Nf6 25. Ree7 Ne8 26. Rb7-/+) 19… Bxd3 (19… Bc4 20. Ne5 Qe4 (20… Bxa2 21. Qe3! (21. Rxd5 Ra7!) 21… Ra7! (21… Qxb2?? 22. Nd3 Qc2 23. Qe6+ Kh8 24. Be5+-) (21… Qe4 22. Qf2! Threatening Re1 trapping the queen. 22… Bb3 23. Re1 Qc2 24. Qh4+- Ra6 25. Rxa6 Nxa6 26. Nd7!) (21… Re8 22. Rc1 Qxb2 (22… Qe4 23. Qd2! again threatens to trap the queen) 23. Re6!+-) (21… a4 22. Rb6+/=) 22. Rc1 Qe4 23. Qd2 Bb3 24. Nd3 Bc4 25. Nc5 Qe2 26. b3 Qxd2 27. Bxd2 Be2 28. Rxd5=) 21. Nxc4 dxc4 22. Qxe4 fxe4 and White’s activity compensates for the pawn deficit.) 20. Qxd5+ Kh8 21. Rc1 Qxb2 22. Bh6! gxh6[] 23. Qe5+ Kg8 24. Qg3+ Kh8 (24… Kf7?? 25. Rc7+) 25. Qe5+=

19. cxd3

White has a winning advantage in all variations.

19… Rd8 20. Bxd6 g6

20… Rxd6 21. Qxf5 and White forces mate. The longest variation is nine moves: 21… h5 22. Qf7+ Kh7 23. Qxh5+ Kg8 24. Re8+ Qxe8 25. Qxe8+ Kh7 26. Rf8 Kh6 27. Rh8+ Kg5 28. Qh5+ Kf4 29. Qe5+ Kg4 30. h3#

21. Rfe1 Rxd6

21… Nd7 22. Qh3+-

22. Re8+ Kg7 23. R1e7+ Kf6 24. Qf4 h6 25. Qxh6 Black resigns. 1-0

Here I missed a checkmate in two moves starting with Qh4+. This is ironic, because I always make my students work on mate in two problems for homework. I think this move is explained by the fact that the position is already winning, so it is more practical do just play a win when one sees it. But better behavior is to take a bit of extra time, especially in an important tournament situation as this was, and make sure that my move filter was screening the most accurate moves possible. If I can manage to consistently adhere to this sort of behavior in the future, it should manifest itself in an overall improvement in strength, and I’ll be able to convert more wins in those positions where I occasionally drop half points to sloppy technique.