Matrix Chess IV

For those who haven’t been following, “Matrix chess” is my name for a training technique where I play a blitz (10-minute) game against the computer but I am allowed to take a time out at one point in the game for as long as I want. It gets its name from the Matrix movies, where Neo has the ability to slow down time in his battles against the computer enemies. The previous posts I wrote on Matrix chess were:

• Trump Cards (in Chess!)
• Matrix Chess… Reloaded
• A Fascinating Experiment

Two weeks ago I sat down for a game against Shredder with its rating set at 2504. I’ve never beaten it at that high a rating before and expected to get thrashed, but this was one game where I was not very impressed with its play. I was White, the opening was a Nimzovich Defense (1. d4 Nc6) and we got to the following position after 23 moves:

Position after 22. … Bd8. White to move.

FEN:1r1brk2/1p3pp1/p2p1nq1/2pPpN2/4P1P1/4BPKR/1PP3Q1/5R2 w – – 0 30

What do you think about this position? Does White have an advantage? What should I play?

Of course the first thing to notice is that the computer just moved its bishop (from h4) to d8, attacking my a-pawn. So one of the important questions is whether White has to defend the pawn or not. Against a human opponent I would not feel any qualms about sacrificing it. But against a computer, I wasn’t so sure. I thought that this could be a position where it just takes the pawn and grinds me down with its computer-like patience and accuracy. So I decided that now was the time to take my time-out.

Well, the time-out lasted two weeks! I got busy with other things and forgot about the game, until this morning I found the slip of paper on my desk with the position written down. “What position is this?” I wondered. Then I looked at it carefully and remembered.

With infinite time, I decided that the first question to ask was not whether I should defend the pawn or not. It’s too soon to decide that. Instead, I asked one of the questions for strategic thinking that I’ve mentioned several times: “What are my (and the opponent’s) best and worst pieces?”

It’s clear what my worst piece is: the knight on c3, which isn’t contributing in any meaningful way to my attack. The next question is: “Where would you like to put the knight, if you had all the time you want?” That answer is pretty easy, too. There’s a great hole on f5 where it would threaten the pawn on d6, threaten a royal fork on e7, and generally make life uncomfortable for Black’s king because it takes away an important flight square. Yes, f5 would be a fantastic square for the knight. Can we get there? Yes! Quite easily, with Ne2-g3-f5.

Another relevant question is, “What side of the board should I play on?” Again, the answer seems clear. I have all of my pieces (except the knight) on the kingside and I have very dangerous attacking threats connected with the open h-file. Craig Mar is fond of saying that even if you are even or down in material, you can win if you have more material in the critical sector of the board. In this case, White definitely has a preponderance of material on the kingside and so I should be able to make something happen. All of this is a very strong argument against playing defensively with 24. Ra1.

Okay, so I’m leaning toward playing 24. Ne2, but there is another question I need to ask. It’s a little bit of an impatient move, saying that White wants some action now. One of the principles of Mike Splane chess is to take your time if you have the advantage. If you’re contemplating a tactical sequence, such as a sacrifice, ask yourself, “Is this a unique opportunity, where if I don’t seize the opportunity now I won’t get another chance?” After all, maybe if I play 24. Ra1 I can follow up with Ne2-g3-f5, but without risking a pawn.

When I looked carefully, I realized that this was in fact a unique opportunity. If White plays 24. Ra1? Black can play 24. … Bg5!, either trading off his worst piece (remember the first question?) or else greatly improving its prospects. In fact, Black should have played 23. … Bg5 on the previous move, instead of being tempted to win material with 23. … Bd8. Shredder has played too much like a human, and this is my chance to take advantage.

Finally, is the sacrifice sound? Notice that this question, which was once my first question, has now become the last question. The more I looked at the position after 24. Ne2! Bxa5? 25. Ng3!, the more I loved it. Black just has huge problems to contend with. If, for instance, 25. … Rfd8 26. Nf5 Rd7, White can even think about playing 27. Nh4 Qf6 28. g5. Black’s pieces are all discombobulated. It looked as if Black’s most solid defense would be to bring the bishop back to d8, but then notice that Black has wasted three tempi (… Bd8, … Bxa4, and … Bd8) just to win a pawn on a5 that is, for the time being, nearly irrelevant.

In fact, Black is probably better off not taking the pawn and playing 24. … Bg5, as he (it?) should have done last move. But in that case, I have won a useful tempo (Nc3-Ne2) for free. So I played 24. Ne2.

The next few moves went more or less as expected: 24. … Bxa5? (just like a human, Shredder continues with its mistaken conception) 25. Ng3 Rfe8 26. Nf5 Bd8 (yup, expected that) 27. Kg3.

Now Shredder played the nearly incomprehensible 27. … Rb8. This really surprised me, wasting another tempo with a do-nothing move on the queenside when the kingside is burning. In Shredder’s defense, the move I had expected, 27. … Bg5, isn’t any good either. I had planned to play 28. Rxh7, but in fact the straightforward 28. Bxg5 is even better. Whichever way Black takes, White is going to triple up on the h-file. Black will have to play … f6 to create a flight square for the king, and then White can snatch the d6 pawn with advantage.

I continued 28. Rh5, preparing to triple on the h-file and also hoping to entice Black to move his (its?) knight to f6. That is, in fact, what it did: 28. … Nf6 29. Rh3 Kf8 and now I missed my chance to end the game.

Position after 29. … Kf8. White to move.

FEN: 1r1brk2/1p3pp1/p2p1nq1/2pPpN2/4P1P1/4BPKR/1PP3Q1/5R2 w – – 0 30

Of course, the first move I thought of was 30. Rh8+, but then after the forced 30. … Ng8 I thought that maybe I was overrating my attack and missing a golden opportunity to take a free, and very important, pawn on d6. So without much time for analysis, I went ahead and played 30. Nxd6?

If I had played 30. Rh8+ Ng8 31. R1h1, Black would absolutely be in a world of hurt. White has two crushing threats, either 32. R1h7 with the idea of Rxg7 or 32. Qh3 with the idea of Rxg8+ forcing mate. Black basically has to resign or give up lots of material. After 30. Nxd6? Black was able to play on with 30. … Nh5+! 31. Rxh5 Qxd6. Even though White still stands much better. Black has managed to trade off my best attacker, the knight on f5. Isn’t it funny how that piece, formerly my worst piece, turned into my best?

I eventually did win the game, but it was ugly. In fact, I botched it to such an extent that at one point Shredder could have forced a draw by repetition. The first time we got to the crucial position, it played the (only) correct move and I had to go back to the previous position. But then the second time, it decided, “Nah, I don’t want a draw” and played a losing move instead! That struck me as very human-like. In fact, Shredder lost this game by playing too much like a human. Go figure that one out, Oracle!