Reno, round 6 — Wild middlegame, exquisite endgame

After my last post there was a little discussion about why my annotations were so good. Well, just to show that I still have a long way to go as an annotator, I will show you my last-round game from Reno, where I evaluated the position completely wrong both when I was playing it and when I was analyzing it later. It was only when I put the game on the computer that I found out how wrong I was.

I thought this was my best game in Reno, and from a competitive viewpoint I still think so. There were mistakes, but that was because both players had to solve some really hard questions. It is almost two games in one. First there is a wild tactical middlegame, where as it turns out I was lucky to survive. But then there was an abrupt change around move 27 to a quieter position and then to an endgame where it was all about strategy and which minor pieces were better and which ones were worse. I was not able to outplay my opponent in the tactical part of the game, but I did outplay him in the strategic part. That was a bit of a surprise to me, because I usually think of myself as a tactical player.

My opponent was Mark Rand, a high class-A player. I’ll post the entire game here, although it will be a very long post. I thought about making this post another two-parter, but it seems that when I do multiple-part series, I sometimes don’t get around to posting the later parts until a lot later!

Dana Mackenzie – Mark Rand

Far West Open 2009

1. e4 c5 2. f4 d5 3. Nf3 de 4. Ng5 Nf6 5. Bc4 e6 6. Nc3 …

As mentioned in an earlier entry, I get this position much more often than I get the queen sacrifice variation (5. … Bg4?! 6. Qxg4!)

6. … Be7 7. Ncxe4 O-O 8. d3 …

Perhaps 8. c3 is a little better, preparing to castle but also keeping the white-squared bishop’s options open.

8. … Nc6 9. Be3?! Nd5 10. Bd2 …

I don’t think this two-step with the bishop was very good. Either 9. c3 or 9. O-O would have been better. White is definitely lagging in development, and should have been punished.

10. … Nd4

I was actually worried more about 10. … Na5, playing to win the two bishops, but this move is also good.

11. c3 Nf5 12. Qe2 Qc7 13. g3 … (diagram)

13. … Bd7?

I was definitely relieved to see this. I was more worried about 13. … b6, but the computer points out that 13. … b5! is even better. This is definitely a difference between a strong master and an A-player. The strong master sees that 13. … b5 is a way of jump-starting Black’s attack. White cannot afford to capture with 14. Bxb5 Rb8 15. a4 a6 16. Bc4 Rxb2 — Black wins his pawn back, the rook is very strong on the second rank, and Black threatens to sink a knight on e3. If White plays 14. Bb3 instead, then Black in two moves (… b5 and … Bb7) reaches the same setup that it took him four moves to reach in the game (… Bd7, … b5, … Bc6, … Bb7). Subtleties like this are very important!

14. Bb3 Bc6 15. O-O (finally) Rad8 16. Bc2 b5 (diagram)

17. Qf2? …

A very instructive and typical (for me) mistake. The natural and good move here is 17. Rae1, bringing White’s last piece into the game. I was seduced by the idea of playing g4 and wanted to play Qf2 to “prevent” … Nh4, which is not even all that dangerous. I was also seduced by the idea that I was “gaining a tempo” by playing Qf2 with an attack on the c-pawn, therefore not allowing Black time to forestall my g4 thrust.

Here are the cold facts. First, 17. Qf2 is not gaining a tempo, it’s forcing Black to play a move (17. … Bb7) that he wanted to play anyway! Second, it’s far from clear that the queen belongs on f2. It was probably better placed on e2, eyeing both the e-file and the square h5. (For example, until now Black could not even think of playing … Rfe8 because of the reply Qh5.) Third, the attack with 18. g4 is premature. Strong masters don’t play this way. Jesse Kraai doesn’t play this way. Remember his saying that is currently at the top of this blog page?

“Bad players like to play with their ‘pretty’ pieces. It is the mark of good players that they will not go on an adventure before they solve the problem of their bad pieces.”

And here I am, playing with my pretty pieces (the queen and the knights) and embarking on an adventure (18. g4) before I have figured out what to do with my queen rook and queen bishop. No wonder I got in trouble. White will wish a little bit later that he had a rook on e1.

17. … Bb7 18. g4? Nh6 19. h3 Qc6 20. Nf3 f5!

This is great stuff. One of my own maxims is that if your opponent tries to prevent a move, it’s often both good psychology and good chess to play exactly that move. My move 20. Nf3 was intended to put the fear of God in him and keep him from playing … f5. But he does it anyway! Notice, though, that this move would not be possible if I had a queen on e2 and rook on e1, as I should.

21. Ne5 Qc8?!

Black played this very quickly. I was surprised and still am surprised, because it is from the strategic point of view worse than 21. … Qc7, even without doing any analysis. The move 21. … Qc7 combines attack and defense, while the move 21. … Qc8 is purely defensive. Ninety percent of the time, at least, the move that creates some threats is better.

22. Ng5 (! – Mackenzie, ? – Fritz)

Because I was so sure that my opponent had made a mistake, I felt it my duty to punish him. And I remained convinced of this even after doing reams of home analysis! Alas, when I put the position on the computer it showed me that I was simply deluded.

22. … fg 23. Qh4 (! – Mackenzie, ? – Fritz) Bxg5 24. fg …

24. … Rxf1+?!

My whole idea was that after 24. … Nf5 25. Qxg4, Black cannot win the exchange with 25. … Nde3 26. Bxe3 Nxe3 because White has the intermezzo 27. Rxf8+ Rxf8 and I can then play 28. Qe2. True enough — but the computer doesn’t stop there. After 28. … Nxc2 29. Qxc2 there follows 29. … Rf5! winning the g-pawn. And it’s not just the g-pawn that is the problem; after … Rxg5+ White’s king is just going to be wide open to attack. White is busted.

25. Rxf1 Nf5 26. Qh5 … (diagram)

26. … Rf8?

Black must have been very frustrated by now that he has all these knight forks at g3 and e3 yet can’t play any of them. If 26. … Nde3 27. Bxe3 Nxe3 28. Qf7+ Kh8 29. Rf6!! is a spectacular, Fischer-esque shot. (If 29. … gf 30. g6 threatening mate on h7 or g7, and if 30. … hg 31. Nxg6 mate!) If 26. … Ng3 27. Qf7+ Kh8 28. Rf2 seems simplest. Black cannot stop White from opening the h-file and putting his rook on h2, with nasty mate threats. So it is imperative for Black to defend f7. But it turns out that 26. … Qc7! was a better way to do it. But he didn’t see it, and I didn’t either. Again, this was something I didn’t notice until the computer pointed it out. The main point is that after 27. Re1 Nde3! cuts the communication between White’s rook and the knight on e5. Basically, White ends up a pawn down in all lines, without enough compensation.

I think that this may have been hard to see, psychologically, because Black already played … Qc8 and committed himself to that square. It’s a little bit tough to say okay, the queen really belonged on c7. But it did, and still does.

I also think that in a wild and tactical position like this, it’s sometimes easy to overlook simple moves.

27. Qxg4! …

My first good strategic judgement of the game! I realize that Black’s threat of … Ng3 is now too strong, and I back away from my attack.

27. … Nde3?!

This was not really necessary, but as I said above, I think that it was really hard for Black to see all the forking possibilities with his knights and not be able to play them. I think he figured that this was his last chance.

28. Bxe3 Nxe3 29. Rxf8+ Kxf8

Black didn’t want to capture either way, but he had no choice! If he takes with the queen, he loses material after 29. … Qxf8? 30. Qxe6+ Kh8 31. Nf7+ Kg8 32. Nd6+ Kh8 33. Qxe3.

30. Qxf4+ Nf5

The smoke has cleared. In some ways I was glad to have survived the tactical brawl without losing an exchange or a pawn. White now has a very solid position, and the only drawback to it was that I wasn’t sure at this point how to play for a win. I was also getting low-ish on time: 11 minutes for 10 moves. A move per minute is when I start feeling as if I am in time trouble. I decided here that centralizing my king would be a good way to burn up a few moves without making any commitments or creating any weaknesses.

31. Kf2 Ke7 32. Ke2 Qd8 33. Kd2 Qd5 34. Qe4 …

Forcing a queen trade. Objectively I felt my winning chances were just as good in a queenless endgame as they were with queens on the board. Plus, a queenless position is simpler to play when you’re low on time.

34. … Nd6 35. Qxd5 … (diagram)

A big decision. Which way should Black recapture? Actually, I think that either way is okay, but they have different pluses and minuses and neither one of them can be said to be an “easy draw.” After spending a few hours on this endgame, I am pretty confident that it is a draw with best play. But it’s wrong to call it a “drawn endgame,” because White has all the winning chances. He has the better bishop, he has a more active knight, and he has the possibility of an outside passed pawn on the kingside in many lines. The proper assessment of the position is +/= rather than =.

35. … Bxd5

A little bit surprising, but I think this is actually better motivated from a strategic point of view than 35. … ed. Optically, 35. … ed may look better because it repairs Black’s pawn weakness, but it gives Black a bad bishop after 36. d4! The only reason this is playable for Black is that he has the tactical resource 36. … Ne4+. If 37. Bxe4 de 38. dc Ke6 White can push his pawns to c6 or g6 or both, but they just become targets. For instance, 39. c6 Ba8 40. c7 Bb7 41. g6 hg and now the knight will have to fall back. Or if 39. Ng4 Kf5 Black wins back his pawn, and I have to think that his active bishop and kingside pawn majority will give him a draw. Or if 36. … Ne4+ 37. Ke3, I think that Black draws with 37. … Nxg5. White’s pawns on g5 and h3 were unfortunately placed.

All in all, it is safer for Black to try to activate his bishop, because this makes its “bad bishop” status less of a liability. However, it would be wrong to think that the bad bishop has completely gone away.

36. d4 cd 37. Bxh7! …

The only way to play for a win. It creates the possibility of a passed pawn on the kingside.

37. … dc+

I think that the most secure way to get a draw is 37. … Nc4+ 38. Nxc4 dc+ 39. Kxc3 Bxc4. Although Black still has a bad bishop, it has a lot of mobility. A trickier line is 37. … Bxa2 38. cd Bd5 39. h4 a5 40. h5 Ne4+ 41. Bxe4 Bxe4 42. Ke3 Bd5, where Black survives by the skin of his teeth after (deep breath) 43. Kf4 b4 44. Nd3 Bc4 45. Nc5 Be2 46. h6 gh 47. gh Kf6 (Black has been counting on this, but White lands a stunning blow) 48. Nxe6! Kg6! (the h-pawn must be eliminated at all costs) 49. Ke5 Kxh6 50. Nc4 (… a4 must be prevented) 50. … Bb5 51. b3 (again, … a4 must be prevented) 52. Ke6 Kf8 53. d5 Ke8 54. d6 Bc6! (taking away Nb7) 55. d7+ Kd8 (mustn’t sac too soon! 55. … Bxd7+?? 56. Nxd7 a4 57. ba b3 58. Nf6+ and wins) 56. Kd6 Bxd7! (now or never!) 57. Nxd7 a4! =. White cannot take on a4 because his knight can’t stop the b-pawn. So White must play 58. Nc5. Now Black cannot play for a win with 58. … a3?? because the knight gets back in time after 59. Ne6+. So Black, too, has to settle for a draw with 58. … ab 59. Nxb3.

I have presented such a long variation (22 moves!) because I was uncertain until the very end whether White was winning or whether it was a draw! Such a long line, of course, cannot be calculated over the board, and so one should really go into it only as a last resort.

What Black should really be thinking about is: What is the best piece trade for me? Do I want to play a N-vs.-N, N-vs.-B, B-vs.-N, or N-vs.-N endgame? I think that the B-vs.-B endgame in the first line above is probably his best bet. N-vs.-N might be okay, but could backfire because of White’s outside passed pawn. B-vs.-N, as in the long variation above, is a very dangerous proposition for him. I don’t think that my opponent thought about the position this way. I just think he had a gut feeling that a bishop was always better than a knight, but that’s not true if it is a bad bishop!

38. Kxc3 … (diagram)

 

38. … Ne4+?!

During the game I thought this was the losing move. Maybe it wasn’t, but as explained above I think that Black has chosen the wrong piece trade. My vote for best move is simply 38. … a5. It keeps White’s king from penetrating on the queenside, and it also threatens … Bxa2. I could also make a general comment about playing committing moves in a time scramble. In general it’s best to play flexibly in time pressure if you can. Once you trade pieces, you can’t un-trade them.

39. Bxe4 Bxe4 40. Kd4 Bd5?

The computer thinks this is the losing move. Black had to play 40. … Bb1!, a tough move to see on the last move of the time control. If 41. a3 then we get a position like the game, except that Black’s bishop is already on its best diagonal. If 41. Kc5 then we get a variation that I wanted to avoid, where after 41. … Bxa2 42. Kxb5 Kd6 Black’s e-pawn is going to turn into a real threat. White may have winning chances here, but I was not sure.

41. a3 …

As mentioned above, I wanted to avoid giving Black chances by playing 41. Kc5? Bxa2. Also, of course, the more pawns on the board the better my winning chances.

In general, I am very proud of the way I handled the rest of the game. It was very Nimzovichian. I passed up several opportunities to win material, because my number one objective was to restrict my opponent’s mobility.

I showed the game to a couple of local masters, Gjon Feinstein and Elliot Winslow (who is, I believe, an IM, though long since retired). Gjon was very impressed with the endgame; Elliot was not impressed because he considered it to be an elementary win for White. I guess this shows that beauty is in the eye of the beholder. I still think there is something to be said for winning an elementary endgame in the most stylish possible manner.

41. … Kd6 42. h4 Bb3

Notice that the bishop is not able to do anything to slow down White’s pawn advance.

43. h5 Ke7 44. Nc6+ Kf7

45. Ke5 …

Sticking with the Nimzovichian plan of blockade and restriction of the opponent’s activity. I’m sure that 45. Nxa7 also wins.

45. … a6 46. g6+ Kg8

 

47. Ne7+! Kf8

Entombing the king is also not recommended: if 47. … Kh8 48. Kf4! Bd1 49. Kg5 e5 50. h6 e4 51. Nf5! and wins. (But not 51. h7? e3 52. Nf5 e2 53. Nd6 Bb3 =)

48. h6! …

The point of White’s play. Black can’t take the knight because 49. h7 would queen the pawn.

48. … gh 49. Kf6 e5 50. g7+ Ke8

White to play and win.

There is one final subtlety here. White must not queen his pawn with 51. g8Q+, because after 51. … Bxg8 52. Nxg8 e4 Black wins! Alas, White’s king is on f6, the very square White’s knight wants to go to! I saw this back on move 47, and I also saw the correct move:

51. Nf5! …

The threat is Nd6+ followed by Nf7, and Black cannot stop the g-pawn.

51. … Bg8  52. Nxh6! …

I am now willing to let his e-pawn run wild, because I will take on g8, move back to h6, and then queen my pawn (with check!) before he queens his.

52. … Bh7 53. Ke5 Kd7 54. Kd5 Kc7 55. Kc5 resigns

What a battle!