Analysis by Houdini reveals:

33..Qxa3 Is losing
32..Rxd3 and the game is probably a draw

Easiest draw 31..Qd4 and white have no better than the exhange of queens!!

As <Hezam> notes earlier, 26...Qf5! would have been big trouble. 26. Qc2! was indicated as best instead of 26.Ng4?

At move 33, Kasparov argues 33...Rxa3 is strongly met by 34. Rf3! and that 33...d2 was the only way to justify 32...cxd3?! He analyses 32...Rxa3 and 32...Rxd3, saying Black goes into an endgame a pawn down but can hold the Draw by exact defence.

31...Qd4 is not mentioned.

I always found these games curious because Karpov decides to go toe to toe with Kasparov instead of working with the tamer, saner Breyer, or maybe even a Berlin. Instead, he plays the most volatile version of the Ruy possible. And in later games in the opening he plays even riskier, with 17..f5.

My take on this game, and most of the Ruys that Karpov lost back then, is that Karpov was always going for a win, and took too many risks with his K-side defenses.

Funny, Stein said that too: Stein vs Karpov, 1972

And Kasparov had two pieces buried on his own q-side for most of the game.

You get no points for endlessly condemning Black's bishop development in the QID, French, and the Zaitsev Ruy.

Well it seems like a good strategy not to use more piece for defense than necessary. Capablanca said you must try to achieve every task with the fewest possible piece, in order to maximize efficiency and therefore what you can achieve on the board.

In the game you mentioned:

Stein vs Karpov, 1972

Stein could have emerged from the opening with a clear advantage with 15. Bd3. No doubt there are games where it is an excellent idea for black to play the bishop to b7 (more in that in a moment), but this is not be best example - with accurate play white gets a clear advantage. Stein just rushed the attack.

<And Kasparov had two pieces buried on his own q-side for most of the game.>

Here is the position after move 18:

click for larger view

I see something other than white pieces <buried> on the queenside for most of the game.

a) There is white rook that can swing over to g3.

b) There is a white bishop bearing down on h7, after a predictable e5 thrust from white.

c) There is a white bishop lined up on h6

d) The queen has access to a host of kingside attacking squares

e) On the next move, white plays his d2 knight to f3.

f) The e1 rook needs no explanation.

At some point white permits himself the pleasure of winning the b5 pawn as well. That drew the b7 bishop to an even worse square to defend a weak pawn. No matter, white still had plenty of pieces left to run black's king over with.

<Everett: You get no points for endlessly condemning Black's bishop development in the QID, French, and the Zaitsev Ruy.>

Actually, I have no problem with the Queen's Indian defense, and play it sometimes myself. It's my belief that in many <1. e4 openings>, black has no business putting the c8 bishop on b7. I find in many cases he would have been better off simply leaving the bishop on c8 and focusing his efforts elsewhere. Sometimes a way that black benefits with a c8 bishop is he can play something like ....Ra7 and swing the rook over to the kingside. Or, the c8 bishop might chop off a white knight going to f5.

The Petrov is an excellent opening because a lot of times that bishop comes out and finds good work on f5 or g4. I believe the Petrov is the soundest defense black has to 1. e4.

Finally, Kasparov has made the same point. He says that in a lot of positions, you can just count up how many pieces white has attacking the kingside, and compare it to how many black has defending. A bishop on b7 is doing something other than defending.

<I mentioned before that tactical ideas rarely occur in isolation and that the most effective ideas usually involve a combination of factors. Here is a fine example which demonstrates all three elements we have been considering.> [Fork, Pin, Skewer]

<This position is from a crucial game in my 1986 World Championship match against Anatoly Karpov (who is playing Black). At the cost of a piece I have launched a ferocious attack but in the position below it looks as though this attack might have burnt itself out.>

click for larger view

<I am a piece down and must do something drastic or the material advantage will soon tell in Black's favour. For a start my queen is pinned against the king. Although this does not involve a threat to win material it is a problem because if the queens are exchanged then my initiative will evaporate and Black will win. So, I played:>

1. d6+!

<This wonderful move (which I was very pleased to have spotted some time ago) leaves Black with three possibilities, all of which succumb to a tactic along the lines of those we have considered in this chapter:>

<1) 1 ... Kxd6 2. Nxf7+ forks king and queen.>

<2) 1 ... Qxd6 2. Nf5+ again forks king and queen.>

<3) 1 ... Ke6 2. Re8+ with a winning skewer.>

<This is a beautiful idea but not that difficult to find when you arrive in the position. However, without an instinctive understanding of how the basic tactical ideas operate it would be very hard to see this several moves back.>

The final position is an art piece! Perhaps I can make a shirt showing this grand work off to others...

Game featured in "New in Chess - Tactics Training - Garry Kasparov "