Author: Amir Ban
Date: 12:39:51 01/03/98
Go up one level in this thread
On January 03, 1998 at 07:34:38, Jeroen Noomen wrote:
>On January 03, 1998 at 04:20:02, Jeroen Noomen wrote:
>
>>Yesterday the first matchgame between Karpov and Anand took place.
>>A very interesting position occurred after the 25th move:
>>
>>Karpov-Anand (1) Slav Defence
>>------------------------------
>>1. d4 d5 2. c4 c6 3. Nf3 Nf6 4. Nc3 e6 5. e3 Nbd7 6. Bd3 dxc4 7. Bxc4
>>b5 8. Bd3 Bb7 9. 0-0 a6 10. e4 c5 11. d5 Qc7 12. Bc2 c4 13. dxe6 fxe6
>>14. Nd4 Nc5 15. Qe2 Bd6 16. f4 e5 17. Ndxb5! (This appears to be much
>>stronger than Nf5) axb5 18. Nxb5 Qb6 19. Nxd6+ Qxd6 20. fxe5 Qxe5 21.
>>Rf5 Qe7 22. Qxc4 (Karpov was still playing very quickly, so I suppose
>>this was all in his home preparation)
>>
>>22. ... Rc8 First interesting moment. I analysed the game in progress
>> with Rebel 9 and it hesitated between Nfd7 and Ncd7.
>> Rc8 never came up. Is there a program that plays Rc8?
>
>Rebel 9's actual choice was Ncd7. Not Nfd7?, this loses because of Bg5.
>My mistake, actually the choice between Ncd7 and Nfd7 came at the next
>move. I did some testing with The King 2.54 on a Pentium 200 MMX:
>
>It likes 22 ... Ncxe4 (depth = 10), although this seems very risky to
>me.
>Variation: 22 ...Ncxe4 23 Bf4,Rc8 24 Qb5+,Kf8 +0.42 for Black.
>After 22 ... Ncd7!? The King gives 23 Bd2!? with the idea of Bb4. Still,
>the score is only -0.17 for White (depth = 10).
>
>>23. Qb5 Ncd7
>
>Here 23 ... Nfd7 is an alternative. According to The King there follows
>24 Bg5,Qe6 25 Rd1!, score +0.37 in favour of White.
>
>>24. Qxb7 Rxc2
>>25. Bg5 Qd6!! This is what it is all about! Rebel plays Qxe4 or
>> alternatively Rf8, but that seems to fail to save the
>> game. F.e. 25 ... Qxe4 26 Qxe4 Nxe4 27 Re1 and the
>> resulting pin is awsome for Black.
>
>This last statement is correct. After 27 Re1,Rc4 28 b3!,Rd4 29 Rf2! (Or
>Bh4!? Rebel 9),h6 30 Rd2!,Rb4 31 a3,hxg5 32 axb4,Ndf6 33 Rd4 the ending
>should be winning for White.
>
>Still, I didn't find a convincing refutation of Rebel's second choice
>25 ... Rf8!? After f.e. 26 Re1,Qe6! (The King) Black intends Qb6+ or
>Kf7-g8 with a very obscure position. Is there a win somewhere after
>25 ... Rf8?
>
>>26. Qa8+ Kf7! Of course, this is the point! Somehow Anand managed
>> to find something at the board, sidestepping the
>> home-analysis of Karpov! I am very impressed.
>
>The King and Rebel favour 26 ... Qb8 27 Qxb8 Nxb8 28 Bxf6 gxf6. I
>suspect
>this might be better for White, but is it really that much!?
>
>>27. Qxh8 Qd4+
>>28. Kh1 Qxe4
>>29. Rf3 Rxg2!! Beautiful!
>>30. Kxg2 Ne5
>>31. Qxg7+!! A fantastic shot by Karpov, still leading to a better
>> endgame for White.
>
>Rebel and The King play this as well.
>
>>31. ... Kxg7
>>32. Bxf6 Kg6
>
>Perhaps Kh6 is better, but I didnot analyse this deeply.
>
>>33. Bxe5 Qxe5
>>34. Rg1!
>>
>>Now Anand managed to win the White b-pawn, leading to a very difficult
>>endgame. After more than 100 moves Karpov went on to win it.
>
>Actually Anand won the a-pawn, not the b-pawn.
>
>>My question: What do other programs think of 22. ... Rc8, 25 ... Qxe4
>>25 ... Rf8, 25 ... Qd6!! and 26 ... Kf7! ?
>
>Amir, Robert, Thorsten, Chris, Don and others: What do your programs
>think of these interesting positions?
>
>Regards, Jeroen
I looked at this game briefly yesterday. I didn't pay much attention to
Junior's eval because I think these unable-to-castle positions are very
difficult for computers and I prefer the judgement of world-class
humans. Junior was not thinking about 22.Rc8 but I though Anand was
doing the right thing. I did look into 25.Qxe4 long enough to see that
it loses.
Today I looked more, and I think Anand indeed missed 22...Ncxe4! It
takes a pawn, opens the c-file for a later Rc8, creates immediate
back-rank threats for white and most of all creates many threats against
the rook f5. The analysis I have is:
22... Ncxe4 23.Bf4 Rc8 25.Qb3 { 25.Qa4 Bc6 26.Qb3 transposes to the main
line. 25.Qb5+ Kf8 26.Re5 Rc5 black advantage } Bd5 26.Rxd5 Nxd5 27.Re1
0-0 28.Qxd5+ Kh8 29.Bxe4 Rxf4 looks very good for black.
I don't see the point of 25...Rf8 26.Re1 Qe6. This may be an issue of
evaluation. Junior non-chalantly plays 27.a4 and evaluates it +0.7 for
white. At one point it jumps up with 27.e5 but then abandons it. This is
a breathtaking line: 27.e5!? Qxf5 28.exf6+ Kf7 29.Re7+ Kg6 30.Rxg7+ Kh5
31.h3. Who wins now ? I have no idea.
About 26...Kf7: While I respect Anand for playing it, seems no doubt
that this is the losing move. Qb8 of course (-0.50).
Amir
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.