Mexico - Zappa 5.5 v Rybka 4.5- Zappa wins $10,000

This forum is for general discussions and questions, including Collectors Corner and anything to do with Computer chess.

Moderators: Harvey Williamson, Steve B, Watchman

Forum rules
This textbox is used to restore diagrams posted with the fen tag before the upgrade.
Post Reply

Mexico - Who do you WANT to win - not who you think will win?

Poll ended at Wed Sep 19, 2007 3:56 pm

Zappa
13
46%
Rybka
11
39%
I dont care
4
14%
 
Total votes: 28

User avatar
Harvey Williamson
Site Admin
Posts: 6079
Joined: Sun Jul 29, 2007 6:57 am
Location: Media City, UK
Contact:

Post by Harvey Williamson »

There is a video of Rybka v GM Fontaine last night. It is a game Father might have played. The GM did lose it at the end. The funniest thing is Dagh(who operates Rybka in the match v Zappa) sleeping at just over 4 minutes in.

http://www.brightcove.tv/title.jsp?title=1200181019

Best Wishes,

Harvey
User avatar
Mark Uniacke
Hiarcs Author
Posts: 1458
Joined: Sun Jul 29, 2007 1:32 pm
Location: United Kingdom
Contact:

Post by Mark Uniacke »

joseMfc wrote:Both engines have made some mistakes, but also we have seen very good games, (Zappa's game 3, Rybka's game 8).

But Rybka's errors have been quite astonishing :shock: ... game 4 and game 9.

The end game with different colours bishops is, for humans, very easy to understand ...

Is this a bug in Rybka code or is so difficult for computers?.

Today last game .... how exciting match!
It is a bad eval in the Rybka version that played.

HIARCS 11 and most other chess programs will not play the blunder.

The standard of chess in the match has been incredibly variable from extremely high to extremely low! :shock:

It shows that even today computer chess has a long way to go...
Best wishes,
Mark

https://www.hiarcs.com
User avatar
Mark Uniacke
Hiarcs Author
Posts: 1458
Joined: Sun Jul 29, 2007 1:32 pm
Location: United Kingdom
Contact:

Post by Mark Uniacke »

Anthony C wrote:A black day for the Zappa team. Zappa got a very good position with Black in a Ruy Lopez, got in time trouble, blundered, and lost. After Qd8 Qb8 Qg8, White has the amazing plan of g5-g6, Kh5, and Bxh6, and can win despite being down a full exchange. Black could probably play for a win after e4 instead of Qxb5, or a draw after c3 instead of Qd8.

In the second game, Rybka simply outplayed Zappa, and again got a position where its deep search was able to find some tricks. If you look at the position 5 full moves before Rxc4 it is hard to believe that Black is winning. However in the endgame Rybka traded rooks and despite having an eval of +7 (Zappa only +2 for White) the game was drawn.

I actually think this match showcases the differences between Rybka and Zappa quite well. Zappa's knowledge is in its evaluation and its relatively accurate guesses about positions. Rybka's knowledge is in its intelligent search which lets it see very deeply very quickly. Interestingly most people associate a gradual score rise with a superiority in evaluation, when in fact it usually indicates a superiority in search, as the opponent is making many little blunders. In contrast when one engine has a superiority in evaluation the score will usually rise sharply after a move because suddenly the engine can secure a positional advantage (whatever the term is that the other engine is unaware of).

Two years ago I would have told you that a good evaluation is more important than a good search, but now I am not so sure. At the very least Rybka has proven that even today's engines on big hardware (Zappa) still make many search errors, especially at such quick timecontrols as 60+20.

cheers,

anthony
First congratulations to you and Erdo for Zappa achieving at least a drawn match with one game to go!

Thankfully when the match is over HIARCS can have its x8 test machine back for testing MP improvements! :D

I agree with the view that Rybka's main strength is in its fast well aimed search. Like you I have come to realise that search is king in the land of chess computers once a reasonable level of eval has been reached. (Warning this does not necessarily apply against humans)

Look at it this way, create two versions of the same program, let one version search 1 ply deeper than the other. Play a match and what score would you expect?

I think you will see 65%+-5%

Now convert that score to Elo gain, so you might find its about 100 Elo or more.

Now consider how you can change your evaluation to get an improvement of 100 Elo!? :shock:

Sadly, the breakthroughs in computer chess have mainly come from improvements to the search, whereas the eval improvements have been a more gradual creep.

Food for thought.... :idea:
Best wishes,
Mark

https://www.hiarcs.com
Father
Member
Posts: 29
Joined: Fri Sep 21, 2007 2:34 am

Are Rybka and Zappa better that Top 8 human beings from Mex?

Post by Father »

Terry McCracken wrote:
Steve B wrote:
Terry McCracken wrote:
Anthony C wrote:A black day for the Zappa team. Zappa got a very good position with Black in a Ruy Lopez, got in time trouble, blundered, and lost. After Qd8 Qb8 Qg8, White has the amazing plan of g5-g6, Kh5, and Bxh6, and can win despite being down a full exchange. Black could probably play for a win after e4 instead of Qxb5, or a draw after c3 instead of Qd8.

In the second game, Rybka simply outplayed Zappa, and again got a position where its deep search was able to find some tricks. If you look at the position 5 full moves before Rxc4 it is hard to believe that Black is winning. However in the endgame Rybka traded rooks and despite having an eval of +7 (Zappa only +2 for White) the game was drawn.

I actually think this match showcases the differences between Rybka and Zappa quite well. Zappa's knowledge is in its evaluation and its relatively accurate guesses about positions. Rybka's knowledge is in its intelligent search which lets it see very deeply very quickly. Interestingly most people associate a gradual score rise with a superiority in evaluation, when in fact it usually indicates a superiority in search, as the opponent is making many little blunders. In contrast when one engine has a superiority in evaluation the score will usually rise sharply after a move because suddenly the engine can secure a positional advantage (whatever the term is that the other engine is unaware of).

Two years ago I would have told you that a good evaluation is more important than a good search, but now I am not so sure. At the very least Rybka has proven that even today's engines on big hardware (Zappa) still make many search errors, especially at such quick timecontrols as 60+20.

cheers,

anthony
I'm very glad Anthony that you take the time and give your analysis on the match and individual games.

I wish Vas would do the same! Of course he does have his own forum, still it would nice if he'd give his take on the match here as well.

Best,
Terry
and of course we also are treated to Mark's Comments and analysis here as well

its very special to have a Premiere Elite World programmer commenting on the games of other commercial engines as they are locked in fatal mortal combat with mucho dinero on the line

History In the making regards
Steve
Indeed, Steve!

Waiting for the Rematch Regards,
Terry
Anthony, Harvey and aTerry. :D :D
I would like to know the meaning and diference between a "Good evaluation and search inteligence".
With best respect,
Pablo
User avatar
Harvey Williamson
Site Admin
Posts: 6079
Joined: Sun Jul 29, 2007 6:57 am
Location: Media City, UK
Contact:

Post by Harvey Williamson »

Hi Pablo,

I am sure someone else can explain this a lot better than me. But evaluation when you look at the position on the board and say its = or black is better or white is better and by how much they are better by judging that position.

Then the engine will search and in that search it may find some good tactics and discover that the initial evaluation of the position is wrong.

Some engines search deeper than others and some are more positionally aware.

Best Wishes,

Harvey
Anthony C
Zappa
Posts: 5
Joined: Tue Sep 25, 2007 8:34 pm

Post by Anthony C »

Hi Harvey,

I think you are basically right but you are caught in the trap of "evaluation = position", "search = tactics". Evaluation can see tactics and search can see positional moves.

The classic examples are game 7 and game 9. In game 7 Rybka plays Rc7, a terrible move. Why? Because it relies on its search and its search wasn't seeing the correct lines where white gives the pawns on the queen side in exchange for new passed pawns on the kingside, which turn out to be stronger.

In game 9, Rybka won basically because of a pin on the d-file (Rd6 -> Bd3 -> Rd1). Zappa could never find a way to break this pin. Zappa evaluates this kind of a pin in the evaluation, except in this case it was doing it wrong. Very wrong. 0.3 evaluation error -> lost position.

So I really think search and evaluation are simply different options to try to improve play. Obviously I think eval works better while Vasik thinks search works better. So far in Mexico things are proving to be more or less a tossup: Zappa played very well in games 3, 5, and (mostly) 7, while Rybka played very well in games 2, 8, and (mostly) 9.

anthony
Uri Blass
Member
Posts: 82
Joined: Sun Aug 12, 2007 1:40 pm

Post by Uri Blass »

Anthony C wrote:Hi Harvey,

I think you are basically right but you are caught in the trap of "evaluation = position", "search = tactics". Evaluation can see tactics and search can see positional moves.

The classic examples are game 7 and game 9. In game 7 Rybka plays Rc7, a terrible move. Why? Because it relies on its search and its search wasn't seeing the correct lines where white gives the pawns on the queen side in exchange for new passed pawns on the kingside, which turn out to be stronger.

In game 9, Rybka won basically because of a pin on the d-file (Rd6 -> Bd3 -> Rd1). Zappa could never find a way to break this pin. Zappa evaluates this kind of a pin in the evaluation, except in this case it was doing it wrong. Very wrong. 0.3 evaluation error -> lost position.

So I really think search and evaluation are simply different options to try to improve play. Obviously I think eval works better while Vasik thinks search works better. So far in Mexico things are proving to be more or less a tossup: Zappa played very well in games 3, 5, and (mostly) 7, while Rybka played very well in games 2, 8, and (mostly) 9.

anthony
I can only agree with the opinion that zappa has superior evaluation and that rybka has superior search but I do not agree that you think that the eval work better and I do not agree that Vas think that the search works better.


If we talk about your opinion
The facts are that you spent time on implementing more efficient parallel search than Vas and you know that parallel search is search and not evaluation.

If we talk about Vasik's opinion then rybla clearly made evaluation changes and not only search changes for the match and rybka missed the win in game 9 because of an evaluation change that allowed zappa to trade rooks.

Uri
User avatar
Terry McCracken
Senior Member
Posts: 1300
Joined: Tue Jul 31, 2007 5:49 pm

Post by Terry McCracken »

Uri Blass wrote:
Anthony C wrote:Hi Harvey,

I think you are basically right but you are caught in the trap of "evaluation = position", "search = tactics". Evaluation can see tactics and search can see positional moves.

The classic examples are game 7 and game 9. In game 7 Rybka plays Rc7, a terrible move. Why? Because it relies on its search and its search wasn't seeing the correct lines where white gives the pawns on the queen side in exchange for new passed pawns on the kingside, which turn out to be stronger.

In game 9, Rybka won basically because of a pin on the d-file (Rd6 -> Bd3 -> Rd1). Zappa could never find a way to break this pin. Zappa evaluates this kind of a pin in the evaluation, except in this case it was doing it wrong. Very wrong. 0.3 evaluation error -> lost position.

So I really think search and evaluation are simply different options to try to improve play. Obviously I think eval works better while Vasik thinks search works better. So far in Mexico things are proving to be more or less a tossup: Zappa played very well in games 3, 5, and (mostly) 7, while Rybka played very well in games 2, 8, and (mostly) 9.

anthony
I can only agree with the opinion that zappa has superior evaluation and that rybka has superior search but I do not agree that you think that the eval work better and I do not agree that Vas think that the search works better.


If we talk about your opinion
The facts are that you spent time on implementing more efficient parallel search than Vas and you know that parallel search is search and not evaluation.

If we talk about Vasik's opinion then rybla clearly made evaluation changes and not only search changes for the match and rybka missed the win in game 9 because of an evaluation change that allowed zappa to trade rooks.

Uri
I think you're right Uri. Can anyone here persuade Vas to comment on these matters?

The Other Side of the Story Regards,
Terry

P.S. Good Eval is nothing without Good Search and Good Search is nothing without Good Eval :wink:
User avatar
Harvey Williamson
Site Admin
Posts: 6079
Joined: Sun Jul 29, 2007 6:57 am
Location: Media City, UK
Contact:

Round 10 final score 5.5-4.5 Zappa wins

Post by Harvey Williamson »

[Event "Blitz:10'+10"]
[Site "?"]
[Date "2007.09.27"]
[Round "10"]
[White "Rybka"]
[Black "Zappa Mexico "]
[Result "1/2-1/2"]
[ECO "A30"]
[PlyCount "118"]
[EventDate "2007.??.??"]
[TimeControl "600+10"]

{4MB, ZapM X Black.ctg, ZAPPA} 1. Nf3 {0} Nf6 {B/0 0} 2. c4 {11} e6 {B/0 0} 3.
Nc3 {8} c5 {B/0 0} 4. g3 {8} b6 {B/0 0} 5. Bg2 {10} Bb7 {B/0 0} 6. O-O {8} Be7
{B/0 0} 7. Re1 {8} d6 {B/0 0} 8. e4 {9} a6 {B/0 0} 9. d4 {8} cxd4 {B/0 0} 10.
Nxd4 {8} Qc7 {B/0 0} 11. Be3 {8} Nbd7 {B/0 0} 12. f4 {8} Rc8 {B/0 0} 13. Rc1 {9
} O-O {B/0 0} 14. f5 {10} e5 {0.42/19 289} 15. Nb3 {13} Qb8 {B/0 0} 16. Qe2 {
151} Rc7 {0.39/19 245} 17. Nd2 {554} Rfc8 {0.37/21 0} 18. g4 {(a4) 153} h6 {
0.43/20 233} 19. h4 {12} Nh7 {0.48/19 228} 20. Bf2 {10} Bc6 {0.48/20 498} 21.
Nd5 {131} Bxd5 {0.48/20 1} 22. exd5 {20} b5 {0.49/19 212} 23. b3 {40} b4 {
0.52/19 129} 24. Qe3 {109} Nc5 {0.51/19 43} 25. Ne4 {117} Nxe4 {0.54/19 113}
26. Bxe4 {14} Rd7 {0.58/19 424} 27. Bf3 {(Kh1) 179} Qc7 {0.52/18 122} 28. Kg2 {
(Ted1) 10} a5 {0.61/17 174} 29. Kh3 {(Th1) 371} Qd8 {0.50/18 195} 30. Qd2 {
(Ted1) 11} Rdc7 {0.52/18 125} 31. Rcd1 {(Ted1) 66} Bf6 {0.52/18 136} 32. Re2 {
(Tc1) 134} Rd7 {0.55/17 101} 33. Qd3 {(Tc1) 31} Be7 {0.51/16 27} 34. a3 {
(Dd2) 47} bxa3 {0.45/17 39} 35. Ra1 {43} Qf8 {0.29/18 0} 36. Rxa3 {55} Bd8 {
0.31/18 0} 37. Kg2 {1339} Rb8 {0.30/22 37} 38. Re1 {(Tea2) 109} Rdb7 {
0.23/20 80} 39. c5 {55} Rb4 {0.23/18 8} 40. c6 {(Dd1) 81} Qe7 {0.12/15 41} 41.
Bg3 {(Dd2) 179} Nf6 {0.09/16 32} 42. g5 {194} hxg5 {0.17/18 1} 43. hxg5 {25}
Ne8 {0.00/18 22} 44. Qd2 {(g6) 14} Rxb3 {-0.01/16 37} 45. Rxb3 {9} Rxb3 {
-0.01/16 0} 46. g6 {284} Bb6 {0.00/18 2} 47. Qa2 {27} Rb4 {0.00/17 9} 48. Rb1 {
(Te4) 34} fxg6 {0.00/17 53} 49. fxg6 {10} Qg5 {0.00/18 24} 50. Rxb4 {11} axb4 {
0.00/18 2} 51. Qb1 {(Dc4) 30} Qd2+ {0.00/19 107} 52. Kf1 {411} Qe3 {0.00/19 49}
53. Qf5 {33} Qg1+ {0.00/22 0} 54. Ke2 {9} Qe3+ {0.00/22 0} 55. Kd1 {(Kf1) 30}
Qb3+ {0.00/21 40} 56. Kd2 {(Ke1) 41} Qc3+ {0.00/20 38} 57. Kd1 {8} Qb3+ {
0.00/22 75} 58. Ke1 {11} Qe3+ {0.00/23 37} 59. Kd1 {(Kf1) 10} Qb3+ {0.00/24 94}
1/2-1/2
User avatar
Mark Uniacke
Hiarcs Author
Posts: 1458
Joined: Sun Jul 29, 2007 1:32 pm
Location: United Kingdom
Contact:

Post by Mark Uniacke »

Terry McCracken wrote:
Uri Blass wrote:
Anthony C wrote:Hi Harvey,

I think you are basically right but you are caught in the trap of "evaluation = position", "search = tactics". Evaluation can see tactics and search can see positional moves.

The classic examples are game 7 and game 9. In game 7 Rybka plays Rc7, a terrible move. Why? Because it relies on its search and its search wasn't seeing the correct lines where white gives the pawns on the queen side in exchange for new passed pawns on the kingside, which turn out to be stronger.

In game 9, Rybka won basically because of a pin on the d-file (Rd6 -> Bd3 -> Rd1). Zappa could never find a way to break this pin. Zappa evaluates this kind of a pin in the evaluation, except in this case it was doing it wrong. Very wrong. 0.3 evaluation error -> lost position.

So I really think search and evaluation are simply different options to try to improve play. Obviously I think eval works better while Vasik thinks search works better. So far in Mexico things are proving to be more or less a tossup: Zappa played very well in games 3, 5, and (mostly) 7, while Rybka played very well in games 2, 8, and (mostly) 9.

anthony
I can only agree with the opinion that zappa has superior evaluation and that rybka has superior search but I do not agree that you think that the eval work better and I do not agree that Vas think that the search works better.


If we talk about your opinion
The facts are that you spent time on implementing more efficient parallel search than Vas and you know that parallel search is search and not evaluation.

If we talk about Vasik's opinion then rybla clearly made evaluation changes and not only search changes for the match and rybka missed the win in game 9 because of an evaluation change that allowed zappa to trade rooks.

Uri
I think you're right Uri. Can anyone here persuade Vas to comment on these matters?

The Other Side of the Story Regards,
Terry

P.S. Good Eval is nothing without Good Search and Good Search is nothing without Good Eval :wink:
Terry, your PS is spot on!
Best wishes,
Mark

https://www.hiarcs.com
User avatar
Dylan Sharp
Senior Member
Posts: 2431
Joined: Fri Aug 10, 2007 12:07 am

Post by Dylan Sharp »

Congratulations to Zappa! :D
User avatar
Mark Uniacke
Hiarcs Author
Posts: 1458
Joined: Sun Jul 29, 2007 1:32 pm
Location: United Kingdom
Contact:

Re: Round 10 final score 5.5-4.5 Zappa wins

Post by Mark Uniacke »

Harvey Williamson wrote:[Event "Blitz:10'+10"]
[Site "?"]
[Date "2007.09.27"]
[Round "10"]
[White "Rybka"]
[Black "Zappa Mexico "]
[Result "1/2-1/2"]
[ECO "A30"]
[PlyCount "118"]
[EventDate "2007.??.??"]
[TimeControl "600+10"]

{4MB, ZapM X Black.ctg, ZAPPA} 1. Nf3 {0} Nf6 {B/0 0} 2. c4 {11} e6 {B/0 0} 3.
Nc3 {8} c5 {B/0 0} 4. g3 {8} b6 {B/0 0} 5. Bg2 {10} Bb7 {B/0 0} 6. O-O {8} Be7
{B/0 0} 7. Re1 {8} d6 {B/0 0} 8. e4 {9} a6 {B/0 0} 9. d4 {8} cxd4 {B/0 0} 10.
Nxd4 {8} Qc7 {B/0 0} 11. Be3 {8} Nbd7 {B/0 0} 12. f4 {8} Rc8 {B/0 0} 13. Rc1 {9
} O-O {B/0 0} 14. f5 {10} e5 {0.42/19 289} 15. Nb3 {13} Qb8 {B/0 0} 16. Qe2 {
151} Rc7 {0.39/19 245} 17. Nd2 {554} Rfc8 {0.37/21 0} 18. g4 {(a4) 153} h6 {
0.43/20 233} 19. h4 {12} Nh7 {0.48/19 228} 20. Bf2 {10} Bc6 {0.48/20 498} 21.
Nd5 {131} Bxd5 {0.48/20 1} 22. exd5 {20} b5 {0.49/19 212} 23. b3 {40} b4 {
0.52/19 129} 24. Qe3 {109} Nc5 {0.51/19 43} 25. Ne4 {117} Nxe4 {0.54/19 113}
26. Bxe4 {14} Rd7 {0.58/19 424} 27. Bf3 {(Kh1) 179} Qc7 {0.52/18 122} 28. Kg2 {
(Ted1) 10} a5 {0.61/17 174} 29. Kh3 {(Th1) 371} Qd8 {0.50/18 195} 30. Qd2 {
(Ted1) 11} Rdc7 {0.52/18 125} 31. Rcd1 {(Ted1) 66} Bf6 {0.52/18 136} 32. Re2 {
(Tc1) 134} Rd7 {0.55/17 101} 33. Qd3 {(Tc1) 31} Be7 {0.51/16 27} 34. a3 {
(Dd2) 47} bxa3 {0.45/17 39} 35. Ra1 {43} Qf8 {0.29/18 0} 36. Rxa3 {55} Bd8 {
0.31/18 0} 37. Kg2 {1339} Rb8 {0.30/22 37} 38. Re1 {(Tea2) 109} Rdb7 {
0.23/20 80} 39. c5 {55} Rb4 {0.23/18 8} 40. c6 {(Dd1) 81} Qe7 {0.12/15 41} 41.
Bg3 {(Dd2) 179} Nf6 {0.09/16 32} 42. g5 {194} hxg5 {0.17/18 1} 43. hxg5 {25}
Ne8 {0.00/18 22} 44. Qd2 {(g6) 14} Rxb3 {-0.01/16 37} 45. Rxb3 {9} Rxb3 {
-0.01/16 0} 46. g6 {284} Bb6 {0.00/18 2} 47. Qa2 {27} Rb4 {0.00/17 9} 48. Rb1 {
(Te4) 34} fxg6 {0.00/17 53} 49. fxg6 {10} Qg5 {0.00/18 24} 50. Rxb4 {11} axb4 {
0.00/18 2} 51. Qb1 {(Dc4) 30} Qd2+ {0.00/19 107} 52. Kf1 {411} Qe3 {0.00/19 49}
53. Qf5 {33} Qg1+ {0.00/22 0} 54. Ke2 {9} Qe3+ {0.00/22 0} 55. Kd1 {(Kf1) 30}
Qb3+ {0.00/21 40} 56. Kd2 {(Ke1) 41} Qc3+ {0.00/20 38} 57. Kd1 {8} Qb3+ {
0.00/22 75} 58. Ke1 {11} Qe3+ {0.00/23 37} 59. Kd1 {(Kf1) 10} Qb3+ {0.00/24 94}
1/2-1/2
Congratulations to Anthony and Erdo for winning the match with Rybka.

And to Harvey for providing the hardware for Zappa and helping Zappa prepare on HIARCS resources (grrrr) :wink:

Great job Harvey also providing the transmission on Playchess and even helping the dark side do their online updates! :wink:

Overall a great performance by Zappa to pull off this match win.
Best wishes,
Mark

https://www.hiarcs.com
User avatar
Ted Summers
Member
Posts: 269
Joined: Tue Jul 31, 2007 2:20 pm
Location: Marietta, GA
Contact:

Post by Ted Summers »

Congratulations to Team Zappa, You guys just made my day. That was one for the record books. Truly a TEAM efford! Well done guys. Thanks also to the Hiarcs Forum for the coverage of this match. It was very exciting! :lol: A BIG thanks to Harvey for providing the hardware also!
User avatar
Mark Uniacke
Hiarcs Author
Posts: 1458
Joined: Sun Jul 29, 2007 1:32 pm
Location: United Kingdom
Contact:

Re: Round 9

Post by Mark Uniacke »

[fen]7B/5k2/4p3/5pR1/4r1p1/8/3Kb3/8 b - - 0 71[/fen]

I just tried little Palm Hiarcs on this position and running on my handheld it knew Ke7 was the correct move and 71...f4??? was a blunder.

The moral: next time Rybka plays a match they need Palm Hiarcs as an advisor :wink:
Best wishes,
Mark

https://www.hiarcs.com
User avatar
Sebastian Boehme
Member
Posts: 264
Joined: Sun Jul 29, 2007 10:59 am
Location: Bernburg (Germany)
Contact:

Post by Sebastian Boehme »

Amazing match,

just got back home and Harvey told me Zappa won.
Finally someone broke the Rybka dominance a little bit.

It feels good to know we could get back the times of before December 2005 sooner or later.
Zappa, Hiarcs, Junior and maybe Shredder and Fritz or some other talented programs will make engine life more equal.

It was too boring to see Rybka rule for over 2 years now. Cheers on a good future and by the way:

Image

A new fish burger for the autumn- Enjoy! :D

Regards,
Sebi
Post Reply