On this webpage GM Larry Kaufman claimed (at least up to 2012) that the following move was brilliant:

[Title "Kasparov vs Shirov (Horgen 1994)"]
[FEN "rnbqkbnr/pppppppp/8/8/8/8/PPPPPPPP/RNBQKBNR w KQkq - 0 1"]
[StartPly "33"]

1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 Nc6 6. Ndb5 d6 7. Bf4 e5 8. Bg5 a6 9. Na3 b5 10. Nd5 Be7 11. Bxf6 Bxf6 12. c3 Bb7 13. Nc2 Nb8 14. a4 bxa4 15. Rxa4 Nd7 16. Rb4 Nc5 17. Rxb7? Nxb7 18. b4 Bg5 19. Na3 O-O 20. Nc4 a5 21. Bd3 axb4 22. cxb4 Qb8 (22... Kh8 23. Ncb6 Ra7 24. Bb5 f5 25. exf5 Rxf5 26. O-O Bh6) 23. h4 Bh6 24. Ncb6 Ra2 25. O-O Rd2 26. Qf3 Qa7 27. Nd7 Nd8 28. Nxf8 Kxf8 29. b5 Qa3 30. Qf5 Ke8 31. Bc4 Rc2 32. Qxh7 Rxc4 33. Qg8+ Kd7 34. Nb6+ Ke7 35. Nxc4 Qc5 36. Ra1 Qd4 37. Ra3 Bc1 38. Ne3

Lichess Stockfish 10+ at depth 20 disagrees with Black's play, claiming that 22... Qb8 is a mistake, and it seems that Kh8 is one of the optimal moves at that point and it eventually leads to a draw. So was GM Larry wrong?

  • 1
    And of course Lichess SF10+ at depth 20 disagrees with this so-called 'brilliant move', suggesting Nce3 instead.
    – user21820
    Apr 12, 2023 at 13:38
  • 2
    chess.com has done so much damage to chess with their attempt to redefine the meaning of "brilliancy"
    – David
    Apr 13, 2023 at 13:52
  • 1
    Depth 20 (what nodes / sec??) may be not be very accurate. A quick check with a browser based Lichess SF 14+ NNUE at depth 42 (~8000k nodes/sec) gives the principal variation starting with 22..Ra7 (eval = +0.1) which continues 23. O-O Kh8 24. Qc2 g6 25. Ncb6 Qe8 26. Nc7 Qd8 27. Ncd5 Qe8 28. Nbd5 Qb8 29. Nb5 Ra6 in a 0.0 eval territory. Of course, the engine overlord's moves are just that, moves, and must be interpreted into human ideas, but 22...Kh8 does not appear to be the best move. Apr 13, 2023 at 14:44
  • 2
    If you look a few moves deeper, the alternative line you gave goes 17. Nce3 0-0 18. Rxb7. And it almost certainly transposes to the other line with optimal play, or you get a position where the computer evaluation is almost identical. Either way, sacrificing this exchange is one of the demands of the position if white wants to maintain the edge.
    – rajb245
    Apr 14, 2023 at 20:04
  • @rajb245: What if I don't buy your 18. Rxb7? What about 17. Nce3 O-O 18. Rb6? How will Black respond? What is your line?
    – user21820
    Apr 20, 2023 at 10:17

1 Answer 1


Because it is brilliant

No, Larry was not wrong, and neither are all the grandmasters and chess commentators who praised (and are still praising) the move 17.Rxb7!!

Nobody, and I am sure neither Kasparov nor Kaufman, is pretending that the move would be winning outright (after all, we are just getting out of the opening and Black hasn't made anything horrendous), so it is quite expected that a deep analysis can reveal a path to equality for Black - especially computer-assisted analysis with engines released a quarter of a century after the game has been played.

But that's not the point.

The concept of sacrificing an exchange for no immediate reward but the long term domination of light squares and the passivity of Black minor pieces is brilliant. Notice how both Bf6 and Nb7 are limited by the pawn structure c3-b4 vs d6-e5, and the lack of open files for Black's rooks.

It made Black's position extremely hard to play and even a top grandmaster like Shirov collapsed in the next 20 moves. And even if Shirov had chosen 22...Kh8 followed by ...f5 with some feeble counterplay, nobody can tell if he would have been able to save or even overturn the game - the position remains extremely delicate to handle. (Btw, I don't understand the move 24.Bb5 in your engine line.)

Producing such a deep and surprising idea demonstrates Kasparov's superior understanding of the position, as he had to balance strategic factors, Black's material gain vs White's blocade, whose effects won't become concrete before at least ten more moves.

Moreover, there is no doubt that White wouldn't be winning either after any other 17th move.

All in all, 17.Rxb7!! is a brilliant concept, and the best move in the position both on a theoretical and a practical sense. There is no doubt that given the chance, any grandmaster would happily go for it nowadays with White - but Black is now avoiding that 12...Bb7 line.

  • also after b4 now black has problems with getting the knight back into the game?
    – cmgchess
    Apr 12, 2023 at 16:55
  • @cmgchess : absolutely, I have added a sentence about it.
    – Evargalo
    Apr 13, 2023 at 6:55
  • Why "there is no doubt that White wouldn't be winning either after any other 17th move"? This statement implies that you can win as Black if White plays Nc3 as I suggested in a comment. Have you succeeded in doing so against SF? Other than that statement, I get what the rest of your answer is saying. The webpage I cited gave me the impression that White is outright winning after that 'brilliant move', but you're saying that it's not about winning.
    – user21820
    Apr 20, 2023 at 10:11
  • @user21820 : No, this statement doesn't imply that Black would be winning after 17.Nce3. It means that Black isn't losing, and with perfect play (something I'm not pretending to be capable of, with or without Stockfish) the game will certainly be drawn.
    – Evargalo
    Apr 20, 2023 at 10:59
  • @Evargalo: Ok... Then I feel your "no doubt" is misleading. I have a very big doubt about the so-called "brilliant move". What if eventually it is shown that Black has a winning strategy after that? Then it wouldn't be brilliant, but just lucky...
    – user21820
    Apr 20, 2023 at 12:44

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.