In the Fray

I’d like to show you my longest and hardest-fought game from last weekend’s tournament in Reno. I have decided that I won’t give a lecture on it, partly because it’s too long and partly because there are too many mistakes. But it is still a great fighting game, and I was proud of winning it.

My opponent was Francisco Achondo, whom I had talked with many times before but never actually played. I had a general idea of the style of chess he liked: uncompromising attack. And this game didn’t disappoint. He played an objectively unsound pawn sacrifice in the opening, but managed to create enough complications to fluster me and cause me to make mistakes.

Analyzing a game like this with the computer is SOOOO different from experiencing it in the heat of combat. The computer calmly and clinically points out all the things we did wrong — both of us. What the computer can’t show you is the psychological pressure we were under.

For me, the pressure was having to defend for about 50 moves in a row, which I am not very good at, and having to cope with the clock. Besides playing super-aggressively, Francisco also plays super-fast. I have dealt with this sort of player before (see Tortoise and Hare). Usually, Hare’s impatience will end up costing him, but Tortoise may have to wait for a very long time for that reward.

Here’s a PGN of the game with very detailed annotations. I’ll show some of the critical positions below.

Dana Mackenzie — Francisco Achondo

I have just played 11. Be2. So far the game has gone very well for me. Black played the Albin Counter Gambit and followed up with the very ambitious 5. … f6, which is basically saying, “I don’t care about winning back the pawn, I’m going to checkmate you.” But in this position I don’t think Black even has microscopic compensation for the pawn. Francisco must have felt that way too, because he decided to offer a second pawn here with 11. … d3?!

I’ve said before that it’s better to sacrifice two pawns than one, so from a practical point of view I think this was a good idea for Black. However, a better way to go about it would have been 11. … O-O-O 12. Nxd4 Bxe2 13. Nxe2, with some pressure on the d-file, and White’s pieces are less organized than they were during the game.

This was the high-water mark of the game for me. White’s position will not look as good as this for another 45 moves. It was also the one point in the game when I got a little bit too cocky. I thought that nothing could go wrong for White in this position, and I flicked out the move 17. Bf5+?! without even thinking about it very hard. It just seemed automatic. I’m two pawns up, so all trades are good.

But the computer considers this a significant mistake. And when you think about it, the computer is right. Why should White lose a tempo in order to trade off one of his best pieces? Instead I should just play 17. h3 or 17. f3 to chase off Black’s bishop, then castle long. With the white-squared bishop available for both defense and attack, White’s position will be better than it was in the game.

Resist stereotypical thinking!

Now comes a really bad decision. Obviously White wants to take back on f3. Which way is correct?

For some reason I “played it safe” here and took with the knight: 26. Nxf3? I guess I was worried that after 26. gf Black would kick my knight with 26. … c5 and then my f-pawn would be weak. But after 26. … c5 27. Nc2 Black can’t take yet because of his loose rook on h8, and that gives me time to play 28. f4 next move. Now, admittedly, I will still have some work to do; it won’t be easy for me to advance those passed pawns. But it’s still way better than the position after 26. Nxf3, where White has voluntarily created two new weaknesses, the pawns on e3 and g2, to go with the existing weak pawn on c4.

After the game I wondered, “How did Francisco manage to get so much compensation for his two-pawn sac?” The answer is that I inflicted some of the pressure on myself, as a result of poor decisions like 26. Nxf3.

Now comes a sequence of moves that looks very embarrassing to both of us when you look at it on a computer. Rybka thinks that every move is a mistake! This is where the “heat of combat” really makes itself felt. We, of course, didn’t know we were playing bad moves; the position is complicated, the pressure is intense, and we’re just doing the best we can.

First, my 29. R7d3?! was not too great, an unforced retreat. Then Francisco played 29. … Qf7?!, also an unforced retreat. This move is so uncharacteristic for him that it needs some explanation. As I’ve said, he typically moved very rapidly, and here his haste came back to haunt him. He picked up his queen with the intention of moving to g4. Then he realized that move would hang a rook, and he tentatively placed the queen on f5 instead. But then he probably saw that 29. … Qf5 30. Nd4 allows White a favorable knight trade. So he finally gave a sigh and put the queen back on f7.

Emboldened by this turn of events, I played 30. b4?, also a mistake. Once again it had more to do with psychology than rational analysis. I was so tired of defending that I did not want to play 30. b3 and defend some more. But that is what White needs, to solidify his position.

After 30. b4? the weak pawn on c4 can no longer be defended by a pawn, and Black should take advantage of this right away with 30. … Re4! Then White has to play either 31. Rd5 or 31. Nd2 (I was intending the latter), and on either one of these moves White’s position starts to look very disharmonious. (See PGN for more details.)

Instead Francisco focused on the wrong target with 30. … Rg6? This would have been good on the previous move, but now the stakes have been raised. He has either overlooked or underrated my counterattack 31. b5! Not only does this chase Black’s knight away from the defense of d8, as in the game, but it also has to give up the defense of e5, which has become a forking square!

After 31. b5! Black can fight back with 31. … Rg4, targeting the c4 pawn again, but now it’s too late. White plays 32. Rd5 Ne7 33. Rd8+ Nc8 and now 34. Ne5 wins the exchange.

But I didn’t see the idea of the knight fork on e5. My time was getting low, and I quickly banged out 31. Rd7? In time pressure I’m just trying to make aggressive moves, but I’m not thinking about the consequences. After 31. … Qe6 32. b5 we’re at another turning point in the game.

Here Black missed a chance to equalize! Once again he should threaten my c-pawn with 32. … Rg4. Now the computer will tell you that this doesn’t equalize, that White retains a significant advantage with 33. R1d5. However, I can tell you for a fact that I wasn’t planning to play that move. I was intending to play 33. Nd4?, because I was so enamored of the trap 33. … Qxd7 34. Nxc6+, which wins Black’s queen and justifies the whole b4-b5 idea in a pretty way. The only problem, however, is that Black can simply play 33. … Nxd4! White can’t recapture with the e-pawn because the rook at d7 would hang (another reason that 31. Rd7 was an unfortunate move). Instead I have to play 34. R1xd4 Rxg2, and according to Rybka White’s advantage is completely gone.

Of course Francisco couldn’t have known that I was planning to blunder after 32. … Rg4. However, it was the best  move to play anyway. I think he just underestimated the importance of pressuring the weak c-pawn.

But in fairness to him, this is only obvious when you look at the game with a computer. In the heat of battle, with no computer to help you, and with lots of difficult choices, it’s not so obvious.

Instead Francisco played a very reasonable idea: 33. … Qxe3+ 34. Qxe3 Rxe3 35. cb Rxc6+ 36. Kd2 Rxa3, giving up the knight for three pawns and obtaining a dangerous-looking pair of passed pawns on the queenside.

But here he got kind of unlucky. After stumbling through the middlegame, I unexpectedly played the endgame perfectly! I’m amazed at this myself. After all the mistakes that Rybka found in my previous moves, I was expecting it to tell me I made all sorts of mistakes in the endgame as well. But in fact it tells me that I was in control the whole way. Funny, it didn’t feel that way! Let’s take a look at the position after the time control.

Now I finally had time to think again, and most importantly make a plan. The first step is to trade a pair of rooks, so I played 41. Nc2! Rb1! (more effective than 41. … Rb2, which accomplishes nothing after 42. Kd3) 42. Ra3+ Rxa3 43. Nxa3. After 43. … Rg1 44. Rg8 b5 we come to the next big decision.

Here I played the counterintuitive (but, according to Rybka, correct!) 45. Nc2! During the game I was not at all sure that this was right, because it lets Black’s queenside pawns start rolling. However, my rationale was that the knight on a3 is my least effective piece. I need to bring it to e1 or e3 to protect my g-pawn. That in turn will free my rook to take his h-pawn and then to defend against the b- and c-pawns. But will I have time to do all of this? I really wasn’t sure.

I think the computer’s point of view is also interesting here. To a human, Black’s b- and c-pawns look quite intimidating. But to Rybka, they are not threatening at all. In fact, there is nothing that White would like better than to lure the pawns forward to b3 and c4, or to c3 and b4, and then blockade them. Because White has more material and can cut Black’s king off, this would surely be death for the pawns.

So, although I was sweating bullets here, I actually played the right move. The game continued 45. … c5 46. Ne3 (46. Ne1 is also possible but just barely — see the PGN if you’re curious) b4 47. Kc2 Re1 48. Nd5! (another accidentally good move, as you’ll see in a second) Re5 49. Nf4!

Hey, ma! Look what I found! The ideal square for my knight!

On f4 the knight can never be dislodged (if Black attacks it, White can play g3), and it controls the critical squares g2 and e2 as well as h5 (which means Black’s h-pawn is a dead duck). And at a moment’s notice it’s free to jump back to d5, which as we’ll see turned out to be very important.

This position makes me think about schematic thinking in the endgame. When you looked at the position after move 44, did you realize that the knight’s eventual destination would be f4? I didn’t! I just kind of lucked into it. In fact, when I moved my knight to d5 it was actually with the idea that I might sac my knight for the two pawns and try to win the rook endgame. It was only after Black stymied this plan with 48. … Re5 that I went to Plan B, which turned out to be better than Plan A.

Lucky or not, White is now in control of the game. But we have one more interesting tactical wrinkle to look at, which occurred after my 57th move.

On the surface it may appear as if Black has made some progress since the last diagram. The pawns have inched forward a bit and the king has joined them.

However, White has not been idle — he has finally rounded up the h-pawn. And the biggest problem for Black is that he cannot make any more progress. If he moves his king to a3, he gets mated. If he pushes either of the pawns, they get blockaded. (See my comments a few moves ago.) Meanwhile, White is threatening to start advancing his own pawn duo.

The logical solution to these problems for Black is 57. … Re3?, which is in fact what Francisco played. He is now threatening … b3+ followed by … c3, and he is preventing 58. h4. But now I have a little combination that ends the resistance. Do you see it?

I’ll give you a little space in case you want to think about it.

La la la la la …

Got it? The answer is 58. Ra8+ Kb5 59. Rb8+ Kc5 60. Rxb4! In knight endgames, always look for forks!! If Black takes the rook, then 61. Nd5+ wins easily.

It was very satisfying, after defending all game long, to be able to land a knockout blow like this. Francisco soldiered on with 60. … Rf3, but after 61. Rb8 White no longer had anything to worry about. For the rest of the game (including a cute finishing move) see the PGN.

This was a very long game, but I hope you found it entertaining! I admire the way that Francisco played, even though he lost, and I think this style will bring him plenty of victories in the future.