Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / GM Ehlvest - Rybka. Game 1.
- - By Victor Zakharov (*****) [ru] Date 2007-07-05 21:45 Edited 2007-07-09 11:50
[Event ""]
[Site ""]
[Date "2007.7.6"]
[Round ""]
[White "*Ehlvest"]
[Black "*Rybka"]
[Result ""]
[Eco ""]
[Annotator ""]
[Source ""]

1. c4 b6 2. Nc3 e6 3. d4 Bb4 4. e3 Bb7 5. Nge2 Nf6 6. a3 Bd6 7. d5 Na6 8. g3
Qe7 9. Bg2 h5 10. h3 exd5 11. cxd5 Nc5 12. b4 Nce4 13. Qd3 Nxc3 14. Nxc3 h4 15.
g4 a5 16. b5 O-O-O 17. Ra2 g5 18. Rc2 Bxa3 19. Bxa3 Qxa3 20. O-O Kb8 21. f4
gxf4 22. Rxf4 Ne8 23. Rxf7 Nd6 24. Rf4 Rhe8 25. Kh2 Re7 26. Qd4 Rde8 27. e4 Qb4
28. Qf2 Re5 29. Qxh4 Bxd5 30. Nxd5 Rxd5 31. g5 Rxb5 32. e5 Qb3 33. Rxc7 Kxc7
34. exd6+ Kxd6 35. g6 Qe3 36. g7 Qe5 37. Qg3 Rg8 38. Kh1 Rxg7 39. Rf6+ Kc7 0-1
Parent - - By VP (*) [au] Date 2007-07-05 23:01
Victor, can you please provide updates?
I think chess OK server is overloaded, and does not allow any more guests :-(

Thanks!
Parent - By Sesse (****) [no] Date 2007-07-05 23:09
If you need a different server, FICS is relaying. Grab xboard or WinBoard (depending on your OS), connect to freechess.org (should be a choice in the menu), log in as a guest, and use "observe 238".

/* Steinar */
Parent - - By Victor Zakharov (*****) [ru] Date 2007-07-05 23:04 Edited 2007-07-05 23:14
[Event ""]
[Site ""]
[Date "2007.7.6"]
[Round ""]
[White "*Ehlvest"]
[Black "*Rybka"]
[Result ""]
[Eco ""]
[Annotator ""]
[Source ""]

1.c4 b6 2.Nc3 e6 3.d4 Bb4 4.e3 Bb7 5.Nge2 Nf6 6.a3 Bd6 7.d5 Na6
8.g3 Qe7 9.Bg2 h5 10.h3 exd5 11.cxd5 Nc5 12.b4 Nce4 13.Qd3 Nxc3
14.Nxc3 h4 15.g4 a5 16.b5 O-O-O 17.Ra2 g5 18.Rc2 Bxa3 19.Bxa3
Qxa3 20.O-O Kb8 21.f4 gxf4 22.Rxf4 Ne8 *

Sorry, connection was lost. Continued

Rybka evaluates the position as equal
[= (+0.00)]  d=19  22.Qd4 Nh7 23.Qxf4 f6 24.Rfc1 Rc8 25.Na4 d6 26.Rc3 Qa2 27.Rxc7 Rxc7 28.Qxd6 Rhc8 (0:00.56)
Parent - - By Lee Ma Hong (**) [ph] Date 2007-07-06 00:17
game is also being broadcast on playchess server

[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "Ehlvest"]
[Black "Rybka"]
[Result "*"]
[Annotator "HiarcsX"]
[PlyCount "68"]

1. c4 b6 2. Nc3 e6 3. d4 Bb4 4. e3 Bb7 5. Nge2 Nf6 6. a3 Bd6 7. d5 Na6 8. g3
Qe7 9. Bg2 h5 10. h3 exd5 11. cxd5 Nc5 12. b4 Nce4 13. Qd3 Nxc3 14. Nxc3 h4 15.
g4 a5 16. b5 O-O-O 17. Ra2 g5 18. Rc2 Bxa3 19. Bxa3 Qxa3 20. O-O Kb8 21. f4
gxf4 22. Rxf4 Ne8 23. Rxf7 Nd6 24. Rf4 Rhe8 25. Kh2 Re7 26. Qd4 Rde8 27. e4 Qb4
28. Qf2 Re5 29. Qxh4 Bxd5 30. Nxd5 Rxd5 31. g5 Rxb5 32. e5 Qb3 33. Rxc7 Kxc7
34. exd6+ Kxd6 *
Parent - By Lee Ma Hong (**) [ph] Date 2007-07-06 00:39
[Event "?"]
[Site "?"]
[Date "2007.07.06"]
[Round "?"]
[White "Ehlvest"]
[Black "Rybka"]
[Result "0-1"]
[Annotator "HiarcsX"]
[PlyCount "78"]

1. c4 b6 2. Nc3 e6 3. d4 Bb4 4. e3 Bb7 5. Nge2 Nf6 6. a3 Bd6 7. d5 Na6 8. g3
Qe7 9. Bg2 h5 10. h3 exd5 11. cxd5 Nc5 12. b4 Nce4 13. Qd3 Nxc3 14. Nxc3 h4 15.
g4 a5 16. b5 O-O-O 17. Ra2 g5 18. Rc2 Bxa3 19. Bxa3 Qxa3 20. O-O Kb8 21. f4
gxf4 22. Rxf4 Ne8 23. Rxf7 Nd6 24. Rf4 Rhe8 25. Kh2 Re7 26. Qd4 Rde8 27. e4 Qb4
28. Qf2 Re5 29. Qxh4 Bxd5 30. Nxd5 Rxd5 31. g5 Rxb5 32. e5 Qb3 33. Rxc7 Kxc7
34. exd6+ Kxd6 35. g6 Qe3 36. g7 Qe5 37. Qg3 Rg8 38. Kh1 Rxg7 39. Rf6+ Kc7 0-1
Parent - - By Sesse (****) [no] Date 2007-07-06 00:39
Ehlvest resigned in a hopeless situation with very little time left. Can't blame him.

Interesting game -- was going towards very closed, then suddenly everything opened up, Ehlvest did what the engines thought was a small mistake, but could probably have hold on to it in a very complex situation. (33 Rxc7, which Ehlvest played, loses -- the question is if 33 Re2 could have hold on to it or not. It seems complicated and will need a bit of analysis -- and it's hard for a human to actually see without engine support.)

Congratulations to the Rybka team for 1-0 so far; will be very interesting to see the rest of the match. Also kudos to Ehlvest for daring to venture into such positions, and holding Rybka off quite a while until what was possibly the decisive mistake in move 33.

Waiting for game 2 tomorrow...

/* Steinar */
Parent - - By Uri Blass (*****) [il] Date 2007-07-06 01:01
It seems that 33.Re2 could lead to advantage of white

Rybka at depth 19 after 33.Re2

Ehlvest - Rybka
1k2r3/2pp4/1p1n4/pr2P1P1/5R1Q/1q5P/4R1BK/8 b - - 0 1


Analysis by Rybka 2.3.2a 32-bit :

1. =  (0.21): 33...Nd6-c8 34.g5-g6 Rb5-b4 35.Re2-f2 Re8-g8 36.Rf4xb4 a5xb4 37.Qh4-e4 c7-c6 38.h3-h4 Qb3-d1 39.Bg2-f3 Qd1-c1 40.h4-h5
2. ²  (0.64): 33...Qb3-e6 34.Rf4-f6 Qe6-e7 35.Qh4-f4 Rb5-c5 36.Qf4-f3 c7-c6 37.Rf6xd6 Rc5xe5 38.Re2-d2 Qe7xg5 39.Qf3-g4 Qg5xg4 40.h3xg4

(,  06.07.2007)

I am not sure that 33.Rxc7 is losing.
It seems that 35.Rd4+ was the last chance of white

After 35.Rd4+ Kc7 rybka see 36.Qf4+ when the position is not clear because white has dangerous passed pawns
Rybka gives advantage for black but the score is improving for white and it suggests that rybka may be wrong(I only analyzed it for few minutes and more time is needed).
Parent - - By Intagrand (*) [gb] Date 2007-07-06 02:09
I can't find a defence for black after Re2. All variations seem to lead to a win for white.

1k2r3/2pp4/1p1n4/pr2P1P1/5R1Q/1q5P/2R3BK/8 w - - 0 33


[Event "Rybka v Ehlvest II"]
[Site "Internet"]
[Date "2007.07.06"]
[Round "1"]
[White "GM Ehlvest, Jaan (USA)"]
[Black "Rybka"]
[Result "*"]
[ECO "E44"]
[WhiteElo "2629"]
[BlackElo "3100"]
[Annotator "HiarcsX"]
[PlyCount "78"]
[EventDate "2007.07.05"]
[EventType "match"]
[EventRounds "6"]
[EventCountry "NET"]

1. c4 b6 2. Nc3 e6 3. d4 Bb4 4. e3 Bb7 5. Nge2 Nf6 6. a3 Bd6 7. d5 Na6 8. g3
Qe7 9. Bg2 h5 10. h3 exd5 11. cxd5 Nc5 12. b4 Nce4 13. Qd3 Nxc3 14. Nxc3 h4 15.
g4 a5 16. b5 O-O-O 17. Ra2 g5 18. Rc2 Bxa3 19. Bxa3 Qxa3 20. O-O Kb8 21. f4
gxf4 22. Rxf4 Ne8 23. Rxf7 Nd6 24. Rf4 Rhe8 25. Kh2 Re7 26. Qd4 Rde8 27. e4 Qb4
28. Qf2 Re5 29. Qxh4 Bxd5 30. Nxd5 Rxd5 31. g5 Rxb5 32. e5 Qb3 {Following moves are from analysis using Rybka.}

33. Re2 $1 Nc8 (33... Qd3 34. Qg4 $1 $18) (33... Qe6 34. Rf6 Qe7 35. Qf4 Nb7 (
35... Rc5 36. g6) 36. e6 $18 Qg7 37. g6) 34. Qg4 $1 Rbxe5 (34... Qd3 35. Rd4
$18 Qc3 (35... Qg6 36. Rxd7) 36. Qe4 d5 37. Rxd5 Rxd5 38. Qxd5 c6 39. Qxc6 Qxc6
40. Bxc6) (34... Rd8 35. e6 dxe6 36. Rxe6 Rg8 (36... Qd1 37. Qxd1 Rxd1 38. g6)
37. g6 Qd3 (37... Rc5 38. Ref6 Qc3 39. Qe6) 38. Rf7 Rb4 39. Qg5 Qc4 40. Qf5 Rb2
41. g7) (34... Qe6 35. Qxe6 Rxe6 36. h4 (36. Rf7 c6 37. h4 a4 38. Rxd7 a3 39.
Rd3 Ra5 40. Ra2 Raxe5 41. Raxa3 b5 42. Kg3 Nb6) 36... Rg6 (36... c6 $2 37. Bh3
$1 $18) 37. Rf7 Rb4 38. Kg3 Rb3+ 39. Kg4 $18) 35. Rxe5 Rxe5 36. g6 Re7 37. Rf8
Qb4 38. Qxb4 axb4 39. Rf7 b3 40. g7 Rxf7 41. g8=Q b2 42. Be4 Re7 43. Bb1 *

I think GM Ehlvest played extremely well given the conditions. I enjoyed the game and hope that GM Ehlvest can continue to play as strong as this and hopefully convert his strong positions to winning ones :). Good luck Ehlvest!
Parent - - By Roland Rösler (****) [de] Date 2007-07-06 04:00
I´m sure that after 33. Re2! white has a winning position. But I´m also absolutely sure that no human (neither Kramnik nor Topalov) wins this position against Rybka after 33. Re2 (34. Qg4 ist very good!).
Great performance from Jaan! I think, he feels fine after Rxc7.
Parent - By Uri Blass (*****) [il] Date 2007-07-06 06:17
I see no reason to be absolutely sure that no human could win this position against rybka after 33.Re2.

Uri
Parent - By George Tsavdaris (****) Date 2007-07-06 11:07

>I´m sure that after 33. Re2! white has a winning position. But I´m also absolutely sure that no human (neither Kramnik nor Topalov) wins this >position against Rybka after 33. Re2 (34. Qg4 ist very good!).


Yes engines after 33.Re2 say that white is better, so with extreme ease we all say that too, but Ehlvest didn't have the help of a computer to know that and obviously he missed something.

You should notice that after 33.Re2 the e-Pawn is attacked by 2 black pieces so it can be captured. Yes if we analyze it with computers we will see that actually if black captures e-Pawn then he would lose but the line is tricky because Ehlvest had to find 33.Re2 Rbxe5 34.Rxe5 Rxe5 35.Rf8+ Re8 36.Rxe8 Nxe8 37.Qe4! that wins. So he had to find the 37.Qe4!.
Easy to see from here when we know there is such a move, but Ehlvest had to analyse all other continuations for move 33 so it was not easy and made the mistake.....

Also as you say even if Ehlvest played 33.Re2 then the chances for a mistake in such position are huge! It's a very complicated position so the probability Ehlvest would make a mistake is very high.
Yet you can't say with absolute confidence as you say that no human would win this against Rybka. Perhaps Ehlvest or another GM would find the correct moves and win this....
Parent - - By plicocf (***) [br] Date 2007-07-06 06:21
I think you are rigth, Uri, the last chance it´s 35. Rd4+,
seemed to be a draw.
I liked the game, very excitant. Interesting, GM Elhvest
was in a better position and could win, but despite the
handicap on time, I think he lost just in the time pressure.
Good game ..... of chess !!! =)

Paulo Soares
Parent - By Hetman (*****) Date 2007-07-06 16:36
Hi,
but it was expected that Rybka will be in worse position from the start. Without opennig book in the phase of the game where the people are relatively better it is usual. The scenario will be repeating itselfes. GM J.E. will be mostly in the better position from the start and the crisis will occur in the middlegame. The better continuations for him will be found.
That what shall be pointed out is good playing by Rybka the endgame. Endgame and opennig are relatively strong side of human GM's.

Regards
Hetman
Parent - By George Tsavdaris (****) Date 2007-07-06 11:16

>After 35.Rd4+ Kc7 rybka see 36.Qf4+ when the position is not clear because white has dangerous passed pawns
>Rybka gives advantage for black but the score is improving for white and it suggests that rybka may be wrong(I only analyzed it for few minutes >and more time is needed).


After some small analysis i think 35.Rd4+ loses.....
Parent - By Milton (***) [us] Date 2007-07-06 12:17
Hi Uri,

A deeper analysis of 33.Re2:

This analysis further confirms your statement that white is probably winning after 33. Re2:

Ehlvest - Rybka
1k2r3/2pp4/1p1n4/pr2P1P1/5R1Q/1q5P/4R1BK/8 b - - 0 1


Analysis by Rybka 2.3.2a mp 32-bit :

=  (-0.01) 33...Qe6 34.Rf6 Qe7 35.Qf4 Nb7 36.e6 Nd8 37.exd7 Qxe2 38.dxe8Q Qxe8 39.g6   00:00:00
=  (-0.01) 33...Qe6 34.Rf6 Qe7 35.Qf4 Nb7 36.e6 Nd8 37.exd7 Qxe2 38.dxe8Q Qxe8 39.g6   00:00:00
=  (-0.08) 33...Qe6 34.Rf6 Qe7 35.Qf4 Nb7 36.e6 Nd8 37.exd7 Qxe2 38.dxe8Q Qxe8 39.g6 Qd7   00:00:00
=  (0.11) 33...Qe6 34.Rf6 Qe7 35.Qf4 Nb7 36.e6 Nd8 37.exd7 Qxe2 38.dxe8Q Qxe8 39.g6 Qe7 40.Qe4   00:00:01
=  (0.00) 33...Nc8 34.g6 Rb4 35.Rxb4 axb4 36.Qe4 c6 37.h4 Kc7 38.h5 Ne7 39.Bh3 Rh8 40.Bg4   00:00:02
=  (0.00) 33...Nc8 34.g6 Rb4 35.Rxb4 axb4 36.Qe4 c6 37.h4 Kc7 38.h5 Ne7 39.Bh3 Rh8 40.Bg4   00:00:05
=  (0.09) 33...Nc8 34.g6 Rb4 35.Rxb4 axb4 36.Qe4 c6 37.h4 Kc7 38.h5 Ne7 39.Bh3 Rh8 40.Bg4   00:00:11
=  (0.15) 33...Nc8 34.g6 Rb4 35.Rxb4 axb4 36.Qe4 c6 37.h4 Qe6 38.Bh3 Qe7 39.Qg4 Qg7 40.h5   00:00:37
=  (0.25) 33...Nc8 34.g6 Rb4 35.Rxb4 axb4 36.Qe4 c6 37.h4 Qg8 38.h5 b3 39.Kg3 Qg7 40.Bf3   00:01:22
²  (0.28) 33...Nc8 34.g6 Rb4 35.Rxb4 axb4 36.Qe4 c6 37.h4 Qg8 38.h5 b3 39.Kg3 Qg7 40.Qd4   00:02:48
²  (0.29) 33...Nc8 34.g6 Rb4 35.Rxb4 axb4 36.Qe4 c6 37.h4 Qg8 38.h5 b3 39.Kg3 Kc7 40.Bf3   00:04:59
²  (0.34) 33...Nc8 34.g6 Rb4 35.Rxb4 axb4 36.Qe4 c6 37.h4 Qg8 38.h5 b3 39.Kg3 Kc7 40.Qf4   00:09:29
²  (0.43) 33...Nc8 34.g6 Rb4 35.Rxb4 axb4 36.Qe4 c6 37.h4 Qc3 38.Qf4 b3 39.Qg3 Qxg3+ 40.Kxg3   00:19:05
²  (0.37) 33...Nc8 34.g6 Rb4 35.Rxb4 axb4 36.Qe4 c6 37.h4 Qc3 38.Qf4 b3 39.Qg3 Qxg3+ 40.Kxg3   00:37:38
±  (0.89) 33...Nc8 34.g6 Rg8 35.Qh7 Qe6 36.Rf6 Qe8 37.e6 dxe6 38.Rexe6 Qd8 39.g7 Rg5 40.Rf7   01:41:44
±  (1.00) 33...Nc8 34.g6 Rg8 35.Qh7 Qe6 36.Rf6 Qe8 37.e6 dxe6 38.Rexe6 Qd8 39.g7 Rg5 40.Rf7   03:41:08

It looks like Rybka likes this position better and better for white, the longer it analyzes it.

Milton
Parent - - By insipid (**) [us] Date 2007-07-06 00:55
Maybe Rybka is right and we're wrong, after all, about blocking the d-pawn.  White actually resigned in a position where black's d pawn is still on d7, don't see that very often now do you?
Parent - - By M ANSARI (*****) [kw] Date 2007-07-06 01:53 Edited 2007-07-06 01:55
I think white went wrong here

1k2r3/1bpp4/1p1n4/pP1Pr3/1q2PRPp/2N4P/2R2QBK/8 w - - 0 29


This is a typical position where a human is doing well but then is punished by the computer for overlooking a tactical shot.  Elvhest here thinks he has a very promising position and once he takes the h4 pawn it looks like his pawn roller is unstoppable so Qxh4? .... but then comes Rybka's Bxd5! and it is all downhill after that.  This shows how terribly frustrating it can be to play against very powerful hardware ... you have to be looking for every possible tactical shot.
Parent - By MoKy (**) [sk] Date 2007-07-06 02:35
[Event "Handicap Match II Crazy Book"]
[Site "?"]
[Date "2007.07.05"]
[Round "?.1"]
[White "Ehlvest, Jaan "]
[Black "Rybka 232a"]
[Result "0-1"]
[ECO "A40"]
[WhiteElo "3100"]
[BlackElo "2660"]
[Annotator "MoKy"]
[PlyCount "78"]

1. c4 b6 2. Nc3 e6 3. d4 Bb4 4. e3 Bb7 5. Nge2 Nf6 6. a3 Bd6 7. d5 Na6 8. g3
Qe7 9. Bg2 h5 10. h3 exd5 11. cxd5 Nc5 12. b4 Nce4 13. Qd3 Nxc3 14. Nxc3 h4 15.
g4 a5 16. b5 O-O-O 17. Ra2 g5 18. Rc2 Bxa3 19. Bxa3 Qxa3 20. O-O Kb8 21. f4
gxf4 22. Rxf4 Ne8 23. Rxf7 Nd6 24. Rf4 Rhe8 25. Kh2 Re7 26. Qd4 Rde8 27. e4 Qb4
28. Qf2 Re5 29. Qxh4 $1 Bxd5 30. Nxd5 Rxd5 31. g5 Rxb5 $2 (31... Rde5 32. g6
Qxb5 33. Rcf2 Rg5 (33... R5e7 34. Rf8 Qe5+ 35. Qf4 Qg7 36. e5 $1 Rxe5 37. Qf6
Qxf6 38. R2xf6 R5e7 39. h4) 34. Rf5 Rxg2+ 35. Kxg2 Nxf5 36. exf5 Qd5+ 37. Kh2
Re1 38. Rg2 Qe5+ (38... Re3 39. g7 Qe5+ 40. Kg1 Re1+ 41. Qxe1 Qxe1+ 42. Kh2)
39. Qg3 Qa1) 32. e5 Qb3 (32... Qb1 33. Re2 Nc8 34. Qh5 Rd8 35. Rf7 Qb3 36. Qf3
Qxf3 37. Bxf3)

1k2r3/2pp4/1p1n4/pr2P1P1/5R1Q/1q5P/2R3BK/8 w - - 0 33


33. Rxc7 $4 (33. Re2 $1 $18 Nc8 (33... Rbxe5 34. Rxe5 Rxe5 35.
Rf8+ Re8 (35... Ne8 36. g6) 36. Rxe8+ Nxe8 37. Qe4) (33... Qd3 34. Qg4 Nc8 (
34... Rbxe5 35. Rxe5 Rxe5 36. Qxd7) 35. Rd4 Qc3 36. Qe4 c6 (36... Qc6 37. Qd3
d5 38. Bxd5) 37. Rxd7 Rb3 38. Qf4) (33... Qe6 34. Rf6 Qe7 35. Qf4 Nb7 (35...
Rxe5 36. Rxe5 Qxe5 37. Qxe5 Rxe5 38. g6 Re7 39. Rf8+ Nc8 40. Rf7) 36. e6 $1 d5
(36... Nd8 37. exd7 Qxe2 38. dxe8=Q Qxe8 39. g6) 37. g6 Rb1 38. Rf7 Qc5 39. Rf2
) (33... Nb7 34. g6 Ka7 (34... Qg8 35. e6 $1 Qxg6 (35... Nd8 36. e7) 36. exd7
Rxe2 37. d8=Q+ Ka7 38. Rg4) 35. g7 Qg8 36. Qf6 Rb4 (36... Nd8 37. Rh4) 37. Rxb4
axb4 38. Rd2) 34. g6 $1 (34. e6 dxe6 (34... Rxe6 35. Rf8 c6 36. Qf4+ Kb7 37.
Rd2 d5 38. g6) 35. Bc6 Rg8 36. Bxb5 Qxb5 37. Rg2 Nd6 38. g6 Kb7) 34... Qd3 (
34... Qe6 35. Rf6 Qe7 36. Qe4 c6 37. Qf4) (34... Qg8 35. Qg4 (35. e6 Qxg6 36.
exd7 Rxe2 37. dxc8=Q+ Kxc8 38. Rf8+ Re8 39. Qg4+ Qxg4 40. Rxe8+ Kd7 41. hxg4
Rh5+ 42. gxh5 Kxe8) 35... Rc5 (35... Rbxe5 36. Rxe5 Rxe5 37. Qxd7 Qe6 38. Qxe6
Rxe6 39. g7 Rg6 40. Rf7) 36. Rf7 d5 37. Ref2) 35. Ree4 Rb2 36. g7 Rg8 37. Qf6
Qd5 (37... Qc2 38. Rg4 Qc5 39. h4 Rf2 40. Ref4 Rxf4 41. Rxf4 Qe3 42. Re4) 38.
h4 Qa2 39. Rg4 Rf2 40. Ref4) 33... Kxc7 34. exd6+ Kxd6 35. g6 Qe3 36. g7 Qe5
37. Qg3 Rg8 38. Kh1 Rxg7 39. Rf6+ Kc7 0-1

MoKy
Parent - - By lkaufman (*****) Date 2007-07-06 02:52
Perhaps Qxh4 was wrong, but Ehlvest told me he definitely did not overlook the reply ...Bxd5, he just judged that the line played was better than anything else he could find.
Parent - By LahamanKM [ph] Date 2007-07-06 04:36
Can you help me if the RYBKA engine can be install on mobile phones?
Parent - - By M ANSARI (*****) [kw] Date 2007-07-06 08:36 Edited 2007-07-06 08:38
Actually after letting Rybka run on the position for a while it seems that white is still OK after Re2 and probably winning ... so Jaan was actually right after all with Qxh4!  I played this position as centaur against Rybka and black loses quickly because Rybka doesn't mind exchanging queens and in the resulting position white pawn passers are much much quicker.  Even if black doesn't exchange queens I don't think black can save the game with best play.

Here is how it could have ended after 33. Re2!

1k2r3/2pp4/1p1n4/pr2P1P1/5R1Q/1q5P/4R1BK/8 b - - 0 33


Nc8 34. g6! Rb4?!

Rybka likes this move even though it seems very dangerous and I would think keeping queens on board would maybe give more chances to save the game.  Allowing exchange of queens makes black position quickly untenable.

35. Rxb4! Qxb4 36. Qxb4 axb4 37. Bd5!

1kn1r3/2pp4/1p4P1/3BP3/1p6/7P/4R2K/8 b - - 0 37


Already here it looks like black is totally lost.  The g pawn is just way too advanced and white controls the g queening square with a very strong Bishop.

.... Ne7 38. Bf7 Rh8 39. g7!

1k5r/2ppnBP1/1p6/4P3/1p6/7P/4R2K/8 b - - 0 39


Rh6 40. Rd2 Kc8 41. Be8 Rh7 42. Bxd7+ Kb8 43. Rg2 c5 44. Be6 Kb7 45. Bd5+ Kc7
46. Rg6 Kd7 47. Be4 Rh5 48. Rd6+ Kc7 49. Re6 Ng8 50. Re8 Nh6 51. Rh8 Ng4+ 52.
Kg2 Rg5 53. g8=Q Rxg8 54. Rxg8 Nxe5 55. h4 +-

I don't know if black could have played better but I don't think he can save the game.  So Uri and Intragrand are correct in pointing out that white had a win had he avoided Rxf7? and played Re2.  But I think Jaan should be given a lot of credit for getting Rybka in a lost position.
Parent - By George Tsavdaris (****) Date 2007-07-06 11:35

>34. g6! Rb4?! 35. Rxb4! Qxb4 36. Qxb4 axb4 37. Bd5!


White has to try much harder for a win after 35...axb4
35...Qxb4? is just very bad.
35...axb4 still gives black a worse position but it's much more complicated....
Parent - - By Vasik Rajlich (Silver) [hu] Date 2007-07-06 07:49
First - this was really an amazing game by Ehlvest. I can't believe how actively and correctly he played the middlegame.

To put this in perspective, since CSVN 2006, Rybka has played in 7 tournaments, totalling about 70 games, many against top programs like Zappa, Hiarcs, Loop, Shredder, and Junior, all running on top hardware, and only twice (Shredder-Rybka, Turino 2006, and Rybka-Spike, Paderborn 2006) did she run into the types of problems she had in this game. I'd put the problems posed by Ehlvest as #2 on this long list.

Some other comments:

1) It's not clear if black can survive after 33. Re2. Anson has used his centaur skills to do a pretty job explaing black's problems here:

http://rybkaforum.net/cgi-bin/rybkaforum/topic_show.pl?tid=1630#pid18205

These are the variations on which black needs to improve.

I haven't looked in really great detail. My first-impression way to try to defend the position would be to get rid of the e5 pawn and try to set up a fortress, for example: 33. Re2 Nc8 34. Qg4 Rbxe5 35. Rxe5 Rxe5 36. g6 Re7 37. Rf8 Qb2 38. Rf7 c6 (no hurry pushing the queenside pawns) 39. g7 Qxf7 40. g8Q Re7. If necessary, the knight can be given up for white's h-pawn.

There are many different version of this defensive idea. I'd give it a 50-50 chance of working.

2) In practice, white's play in these variations needs to be extremely precise. I would like to meet the human who could do it without computer assistance.

3) It seems that Ehlvest underestimated his position, because his 33. Rxc7 leads pretty much by force to an endgame where I am sure that he evaluated his chances as much worse. This brings up an interesting idea for future matches: let the grandmaster see Rybka's evaluation.

4) As good computer chess analysts know, the positions around move 32 and 33 are exactly the sort of positions where computers can make serious misevaluations. Too many things are going on, both sides have passed pawns marching up the board, and the engine must report only a score, not also an "unclearness quotient".

Vas
Parent - - By MoKy (**) [eu] Date 2007-07-06 10:57 Edited 2007-07-06 11:27
After 33. Re2 Nc8 34. g6! black cannot take fort on e5 so g+h pawns go like rolling stones.
Sac on a3 was beginning of great antirybka strategy.
Mario
Parent - - By George Tsavdaris (****) Date 2007-07-06 11:51
Can you give the win after 34. g6 Rb4 35. Rxb4 axb4      ?
White has other tries for 35 move of course like 35.Ree4 or 35.Ref2 but you have to prove one of these win.

Even this way i think 34.Qg4! is the move that wins for white for sure.....
Parent - By MoKy (**) [sk] Date 2007-07-06 13:49 Edited 2007-07-06 14:37
OK 34. g6 Rb4! 35. Rxb4 axb4! 36. Qe4 c6 37. h4 Qc3 +=, 34. Qg4! +- but instead 39. g7 is better 39.h4!
And position after 33. Rxc7? Kxc7 34. exd6+ Kxd6 35. Rd4+ is equal.
Mario
Parent - - By George Tsavdaris (****) Date 2007-07-06 11:48 Edited 2007-07-06 11:53

>I haven't looked in really great detail. My first-impression way to try to defend the position would be to get rid of the e5 pawn and try to set up a >fortress, for example: 33. Re2 Nc8 34. Qg4 Rbxe5 35. Rxe5 Rxe5 36. g6 Re7 37. Rf8 Qb2 38. Rf7 c6 (no hurry pushing the queenside pawns) 39. >g7 Qxf7 40. g8Q Re7. If necessary, the knight can be given up for white's h-pawn.


This fails to 39.Rf8   in which white threatens mate (c8)....

>3) It seems that Ehlvest underestimated his position, because his 33. Rxc7 leads pretty much by force to an endgame where I am sure that he >evaluated his chances as much worse. This brings up an interesting idea for future matches: let the grandmaster see Rybka's evaluation.


So let's add another 300 ELO on GM's favor..... :(
I find this a truly bad idea.... :-)
Parent - By Vasik Rajlich (Silver) [hu] Date 2007-07-07 16:51
The rook just came from f8.

Vas
Parent - - By lkaufman (*****) Date 2007-07-06 12:59
He felt he was much better, but didn't see how to answer 33...Qd3 (after 33 Re2). He would definitely have played Re2 if he realized than Black had to retread the knight, he said.
Parent - - By Sepultura [ar] Date 2007-07-06 19:52
I think the position after 33.Re2 should be analysed by brute force. Like Vas said, that kind of positions,
with lot of passed pawns and others factors going on, are the worst nightmare for any engine .
In my pc, Rybka thinks about 40 kN/s, while Deep Junior 10 does up to 1500 kN/s, so perhaps it would be better
to leave our god Rybka and analyse with others engines... I cant do it because I only got Rybka...just try!
Anyway, theres no doubt it was a very good game of Elvhest
Regards
Parent - By George Tsavdaris (****) Date 2007-07-07 14:42
Analyse by brute force?
This is meaningless if i understood well what you've said....
Parent - By tdh1967 (*) Date 2007-07-07 12:46
I am sure by now this is not news but I have ran the position after 33.Re2 twice on newest two versions of 32 bit Rybka at 40/2 time control and Rybka shows white wins.  A third game with version 2.2mp 32 bit gave the same result at 40/2.
Parent - By Vasik Rajlich (Silver) [hu] Date 2007-07-07 15:51
Up Topic Rybka Support & Discussion / Rybka Discussion / GM Ehlvest - Rybka. Game 1.

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill