Kosteniuk played a serie of RYBKA 3 ONLY moves- Possible?

[Cubeman]

I personally dont believe that she could do that.

Oh!?, but she did ...

are you sure that info ? I don´t see anything that in chessbase web ?

toilet II ? I don´t believe too. I like that GM woman ;*)

Oliver, don't worry.It is just a joke

[Terry McCracken]

In the game Kosteniuk played a series of single MUST play moves. after 22-b5. For this move Hou almost thought for half an hour and Kosteniuk wasnt known for being a super super GM player, no?

Take a look for yourself, play with Rybka 3 in the Fritz 11 online environment without special settings. Then you can see that the woman ALWAYS found the single solution of the analysis of Rybka 3. I personally dont believe that she could do that. Up to Rxh7 she had found 14 moves in whole at best after RYBKA 3. While Hou deviated from the Rybka scales.

Such a serie of only moves is impossible to find. IMO.

Now when I looked Kosteniuk wasnt suddenly able to find the winning b6 move instead played only Qe6. How strange.

Yawn... Let's see the game Rolf.

[Tony Thomas]

In the game Kosteniuk played a series of single MUST play moves. after 22-b5. For this move Hou almost thought for half an hour and Kosteniuk wasnt known for being a super super GM player, no?

Take a look for yourself, play with Rybka 3 in the Fritz 11 online environment without special settings. Then you can see that the woman ALWAYS found the single solution of the analysis of Rybka 3. I personally dont believe that she could do that. Up to Rxh7 she had found 14 moves in whole at best after RYBKA 3. While Hou deviated from the Rybka scales.

Such a serie of only moves is impossible to find. IMO.

Now when I looked Kosteniuk wasnt suddenly able to find the winning b6 move instead played only Qe6. How strange.

Yawn... Let's see the game Rolf.

I am not sure if some people realize it.. Regardless of which game you analyze you will find the top engine agreeing with more moves from the winner/person who had the advantage unless the win/advantage was due to one terrible blunder from the opponent.

[swami]

In the game Kosteniuk played a series of single MUST play moves. after 22-b5. For this move Hou almost thought for half an hour and Kosteniuk wasnt known for being a super super GM player, no?

Take a look for yourself, play with Rybka 3 in the Fritz 11 online environment without special settings. Then you can see that the woman ALWAYS found the single solution of the analysis of Rybka 3. I personally dont believe that she could do that. Up to Rxh7 she had found 14 moves in whole at best after RYBKA 3. While Hou deviated from the Rybka scales.

Such a serie of only moves is impossible to find. IMO.

Now when I looked Kosteniuk wasnt suddenly able to find the winning b6 move instead played only Qe6. How strange.

Yawn... Let's see the game Rolf.

I am not sure if some people realize it.. Regardless of which game you analyze you will find the top engine agreeing with more moves from the winner/person who had the advantage unless the win/advantage was due to one terrible blunder from the opponent.

Ofcourse, Top engines do agree with strong humans on moves from Closed position games with few pieces on the board.

[Uri Blass]

In the game Kosteniuk played a series of single MUST play moves. after 22-b5. For this move Hou almost thought for half an hour and Kosteniuk wasnt known for being a super super GM player, no?

Take a look for yourself, play with Rybka 3 in the Fritz 11 online environment without special settings. Then you can see that the woman ALWAYS found the single solution of the analysis of Rybka 3. I personally dont believe that she could do that. Up to Rxh7 she had found 14 moves in whole at best after RYBKA 3. While Hou deviated from the Rybka scales.

Such a serie of only moves is impossible to find. IMO.

Now when I looked Kosteniuk wasnt suddenly able to find the winning b6 move instead played only Qe6. How strange.

Yawn... Let's see the game Rolf.

Here is most of the game copied from the chessbase site.
No evidence for cheating.

Rybka does not play 28.Qg3 and prefers 28.g4(except fast time control)
and I used slow PIV.


[Event "6'/40+6'/40+6'/40"]
[Site "MyTown"]
[Date "????.??.??"]
[Round "?"]
[White "New game"]
[Black "Rybka 3 1-cpu 32-bit"]
[Result "*"]
[PlyCount "93"]
[WhiteTeam "Netherlands"]
[BlackTeam "Kazakstan"]
[TimeControl "40/360:40/360:40/360"]

{128MB, Junior9.ctg, URI-PC
} 1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 Nc6 5. Nc3
d6 6. Be3 Nf6 7. Be2 Be7 8. O-O a6 9. a4 O-O 10. f4 Qc7 11. Kh1 Bd7 12. Nb3 b6
13. Qe1 Bc8 14. Qg3 Bb7 15. f5 Kh8 16. Rad1 Rae8 17. fxe6 fxe6 18. Qh3 Bd8 19.
Nd4 Nxd4 20. Rxd4 e5 21. Rc4 Qb8 22. Rd1 b5 23. axb5 axb5 24. Nxb5 Nxe4 25. Bd3
Nf6 26. Rh4 e4 27. Be2 Bc8 28. Qg3 Ba6 29. c4 Bxb5 30. cxb5 Bb6 31. Bf4 Qa7 32.
Bxd6 Bf2 33. Qf4 Nd5 34. Qc1 Rc8 35. Qd2 Rfd8 36. Rxh7+ Kxh7 37. Qxd5 Qe3 38.
Bg4 Ra8 39. Qe6 Kh8 40. Qe7 Qh6 41. h3 Qg6 42. Qe5 Bb6 43. Bh5 Qh6 44. Bg4 e3
45. Qe4 Qf6 46. Rd5 Ra1+ 47. Kh2 *

New game - Rybka 3 1-cpu 32-bit
[d]1qbbrr1k/6pp/3p1n2/1N6/4p2R/4B2Q/1PP1B1PP/3R3K w - - 0 1

Analysis by Rybka 3 1-cpu 32-bit :

28.Qg3
± (1.11) Depth: 2 00:00:00
28.Qg3
± (1.15) Depth: 3 00:00:00
28.Qg3
² (0.60) Depth: 4 00:00:00
28.Qg3 Be7
² (0.59) Depth: 5 00:00:00
28.Qg3 Be7 29.Nxd6 Bxd6 30.Qxd6 Qxb2 31.Qc7
² (0.54) Depth: 6 00:00:00 6kN
28.Qg3 Be7 29.Nxd6 Bxd6 30.Qxd6 Qxb2 31.Qc7 Be6
² (0.52) Depth: 7 00:00:00 12kN
28.Qg3 Be7 29.Nxd6 Bxd6 30.Qxd6 Qxb2 31.Qc7 Qa2 32.Bd4
² (0.52) Depth: 8 00:00:01 28kN
28.Qg3 Be7 29.Nxd6 Bxd6 30.Qxd6 Qxb2 31.Qc7 Qa2 32.Bd4 e3 33.Qc5
² (0.53) Depth: 9 00:00:03 68kN
28.Qg3 Be7 29.Nxd6 Bxd6 30.Qxd6 Qxb2 31.Qc7 Qb7 32.Qxb7 Bxb7 33.Bb5 Rd8 34.Rxd8
² (0.49) Depth: 10 00:00:09 216kN
28.g4
² (0.69) Depth: 11 00:00:50 1036kN
28.g4 Bc7 29.c4 d5 30.Nxc7 Qxc7 31.Rxd5 Be6 32.Rd1 Qb7 33.Bd4
² (0.58) Depth: 11 00:00:53 1091kN
28.g4 d5 29.Rxd5 Be6
± (0.75) Depth: 12 00:01:28 1736kN
28.g4 d5 29.Rxd5 Be6 30.Rd6 Qc8 31.c4 Re7 32.Nd4 Bd7 33.Nf5 Bxf5 34.gxf5 Qb7
± (0.88) Depth: 13 00:03:13 3691kN
28.g4 Re7 29.Nxd6 Rd7 30.Nxc8 Rxd1+ 31.Bxd1 Qxc8 32.c3 Qa6 33.g5 Qd3 34.Bb3 Nh5 35.Bg1 Qxh3 36.Rxh3 Nf4 37.Re3 Bxg5 38.Rxe4 Nd3
± (0.90) Depth: 14 00:08:05 9659kN
28.g4 Re7 29.Qg3 Ba6 30.Qxd6
± (1.06) Depth: 15 00:17:41 21105kN

(Uri, MyTown 18.09.2008)

[Uri Blass]

In the game Kosteniuk played a series of single MUST play moves. after 22-b5. For this move Hou almost thought for half an hour and Kosteniuk wasnt known for being a super super GM player, no?

Take a look for yourself, play with Rybka 3 in the Fritz 11 online environment without special settings. Then you can see that the woman ALWAYS found the single solution of the analysis of Rybka 3. I personally dont believe that she could do that. Up to Rxh7 she had found 14 moves in whole at best after RYBKA 3. While Hou deviated from the Rybka scales.

Such a serie of only moves is impossible to find. IMO.

Now when I looked Kosteniuk wasnt suddenly able to find the winning b6 move instead played only Qe6. How strange.

Yawn... Let's see the game Rolf.

Here is most of the game copied from the chessbase site.
No evidence for cheating.

Rybka does not play 28.Qg3 and prefers 28.g4(except fast time control)
and I used slow PIV.


[Event "6'/40+6'/40+6'/40"]
[Site "MyTown"]
[Date "????.??.??"]
[Round "?"]
[White "New game"]
[Black "Rybka 3 1-cpu 32-bit"]
[Result "*"]
[PlyCount "93"]
[WhiteTeam "Netherlands"]
[BlackTeam "Kazakstan"]
[TimeControl "40/360:40/360:40/360"]

{128MB, Junior9.ctg, URI-PC
} 1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 Nc6 5. Nc3
d6 6. Be3 Nf6 7. Be2 Be7 8. O-O a6 9. a4 O-O 10. f4 Qc7 11. Kh1 Bd7 12. Nb3 b6
13. Qe1 Bc8 14. Qg3 Bb7 15. f5 Kh8 16. Rad1 Rae8 17. fxe6 fxe6 18. Qh3 Bd8 19.
Nd4 Nxd4 20. Rxd4 e5 21. Rc4 Qb8 22. Rd1 b5 23. axb5 axb5 24. Nxb5 Nxe4 25. Bd3
Nf6 26. Rh4 e4 27. Be2 Bc8 28. Qg3 Ba6 29. c4 Bxb5 30. cxb5 Bb6 31. Bf4 Qa7 32.
Bxd6 Bf2 33. Qf4 Nd5 34. Qc1 Rc8 35. Qd2 Rfd8 36. Rxh7+ Kxh7 37. Qxd5 Qe3 38.
Bg4 Ra8 39. Qe6 Kh8 40. Qe7 Qh6 41. h3 Qg6 42. Qe5 Bb6 43. Bh5 Qh6 44. Bg4 e3
45. Qe4 Qf6 46. Rd5 Ra1+ 47. Kh2 *

New game - Rybka 3 1-cpu 32-bit
[d]1qbbrr1k/6pp/3p1n2/1N6/4p2R/4B2Q/1PP1B1PP/3R3K w - - 0 1

Analysis by Rybka 3 1-cpu 32-bit :

28.Qg3
± (1.11) Depth: 2 00:00:00
28.Qg3
± (1.15) Depth: 3 00:00:00
28.Qg3
² (0.60) Depth: 4 00:00:00
28.Qg3 Be7
² (0.59) Depth: 5 00:00:00
28.Qg3 Be7 29.Nxd6 Bxd6 30.Qxd6 Qxb2 31.Qc7
² (0.54) Depth: 6 00:00:00 6kN
28.Qg3 Be7 29.Nxd6 Bxd6 30.Qxd6 Qxb2 31.Qc7 Be6
² (0.52) Depth: 7 00:00:00 12kN
28.Qg3 Be7 29.Nxd6 Bxd6 30.Qxd6 Qxb2 31.Qc7 Qa2 32.Bd4
² (0.52) Depth: 8 00:00:01 28kN
28.Qg3 Be7 29.Nxd6 Bxd6 30.Qxd6 Qxb2 31.Qc7 Qa2 32.Bd4 e3 33.Qc5
² (0.53) Depth: 9 00:00:03 68kN
28.Qg3 Be7 29.Nxd6 Bxd6 30.Qxd6 Qxb2 31.Qc7 Qb7 32.Qxb7 Bxb7 33.Bb5 Rd8 34.Rxd8
² (0.49) Depth: 10 00:00:09 216kN
28.g4
² (0.69) Depth: 11 00:00:50 1036kN
28.g4 Bc7 29.c4 d5 30.Nxc7 Qxc7 31.Rxd5 Be6 32.Rd1 Qb7 33.Bd4
² (0.58) Depth: 11 00:00:53 1091kN
28.g4 d5 29.Rxd5 Be6
± (0.75) Depth: 12 00:01:28 1736kN
28.g4 d5 29.Rxd5 Be6 30.Rd6 Qc8 31.c4 Re7 32.Nd4 Bd7 33.Nf5 Bxf5 34.gxf5 Qb7
± (0.88) Depth: 13 00:03:13 3691kN
28.g4 Re7 29.Nxd6 Rd7 30.Nxc8 Rxd1+ 31.Bxd1 Qxc8 32.c3 Qa6 33.g5 Qd3 34.Bb3 Nh5 35.Bg1 Qxh3 36.Rxh3 Nf4 37.Re3 Bxg5 38.Rxe4 Nd3
± (0.90) Depth: 14 00:08:05 9659kN
28.g4 Re7 29.Qg3 Ba6 30.Qxd6
± (1.06) Depth: 15 00:17:41 21105kN

(Uri, MyTown 18.09.2008)

after 28.Qg3? Ba6

New game - Rybka 3 1-cpu 32-bit
[d]1q1brr1k/6pp/b2p1n2/1N6/4p2R/4B1Q1/1PP1B1PP/3R3K w - - 0 1

Analysis by Rybka 3 1-cpu 32-bit :

29.Qxd6
² (0.67) Depth: 2 00:00:00
29.Qxd6
² (0.49) Depth: 3 00:00:00
29.Qxd6
² (0.62) Depth: 4 00:00:00
29.Qxd6 Qxd6 30.Rxd6 Nd5
² (0.63) Depth: 5 00:00:00
29.c4 Bxb5 30.cxb5 Be7
² (0.40) Depth: 6 00:00:00 8kN
29.c4
² (0.60) Depth: 7 00:00:00 14kN
29.c4 Bxb5 30.cxb5 Bc7 31.Bg5
= (0.04) Depth: 7 00:00:01 31kN
29.Qxd6 Qxd6 30.Rxd6 Nd5 31.Rh3 Nxe3 32.Rxe3 Bg5 33.Rc3
= (0.20) Depth: 7 00:00:01 34kN
29.Qxd6 Qxd6 30.Rxd6 Nd5 31.Rh3 Nxe3 32.Rxe3 Bg5 33.Rc3
= (0.20) Depth: 8 00:00:01 39kN
29.Qxd6 Qxd6 30.Rxd6 Nd5 31.Rh3 Nxe3 32.Rxe3 Bg5 33.Rb3 Rf2 34.Rxa6 Rxe2 35.Ra1 Rd2 36.c4 Bf6
= (0.09) Depth: 9 00:00:03 81kN
29.c4 Be7 30.Nxd6 Bxd6 31.Qxd6 Qxb2
² (0.26) Depth: 9 00:00:05 142kN
29.c4
² (0.46) Depth: 10 00:00:08 204kN
29.c4 Bxb5 30.cxb5 Bc7 31.Bc4
² (0.53) Depth: 10 00:00:09 230kN
29.c4 Bxb5 30.cxb5 Bc7 31.Bc4
² (0.48) Depth: 11 00:00:12 317kN
29.c4 Bxb5 30.cxb5 Bb6 31.Bf4 Qa7 32.Bxd6 Bf2 33.Qf4 Bxh4 34.Bxf8 Rxf8 35.Qxh4 Qa2 36.Qg3 Qxb2 37.Qe3 Rc8
² (0.36) Depth: 12 00:00:45 1463kN
29.c4 Bxb5 30.cxb5 Bb6 31.Bf4 Qa7 32.Bxd6 Bf2 33.Qf4 Nd5 34.Qc1 Bxh4 35.Bxf8 Qf2 36.Rxd5 Qxe2 37.Bd6 e3 38.Re5 Ra8 39.Rxe3 Qxb5
² (0.43) Depth: 13 00:01:01 1794kN
29.c4 Bxb5 30.cxb5 Bb6 31.Bf4 Qa7 32.Bxd6 Bf2 33.Qf4 Nd5 34.Qc1 Bxh4 35.Bxf8 Qf2 36.Rxd5 Qxe2 37.Bd6 e3 38.Re5 Rd8 39.Rxe3 Qxb5 40.Re5
= (0.25) Depth: 14 00:01:31 2849kN
29.c4 Bxb5 30.cxb5 Bb6 31.Bf4 Qa7 32.Bxd6 Bf2 33.Qf4 Nd5 34.Qc1 Bxh4 35.Bxf8 Qf2 36.Rxd5 Qxe2 37.Bd6 e3 38.Re5 Rd8 39.Rxe3 Qxb5 40.Re5 Qd3 41.Bc5 h6 42.Bg1 Bf6
= (0.21) Depth: 15 00:04:10 7915kN
29.c4 Bxb5 30.cxb5 Bb6 31.Bf4 Rd8 32.Qh3 Qc7 33.Bg5 Qc2 34.Bxf6 Rxf6 35.Rxh7+ Kg8 36.Rh8+ Kf7 37.Bh5+ Ke7 38.Qg4 Qxd1+ 39.Qxd1 Rxh8 40.h3 Rf5 41.Bg6 Re5 42.Qg4 e3 43.Bf5 Rf8
= (0.00) Depth: 16 00:52:47 73030kN

(Uri, MyTown 18.09.2008)

[Rolf]

Ofcourse, Top engines do agree with strong humans on moves from Closed position games with few pieces on the board.

a) It was a highly tactical position with K-attack, no fortress, nothing

b) I didnt know that a 2500 woman could suddenly play ice-cool like a Rybka with 2700+. Unfortunately we dont have the times anymore, but I followed the game live and after 22-b5 try to imagine how someone lower than a super GM could play all these moves with White. I left after she also found 36.Rxh7! and wrote my message as the opener of this thread. To Uri's true mention of better g4 than Kg3 I made a seperate message to Uri.

Please try to take such messages in context. That wasnt my deep judgement with 100% certainty, it was what I would have done in the tournament hall too. You watch a game, look at your display and wonder how this woman could always play the ONLY winning move compared to the one or two or even all the rest of LOOSING moves. You become somewhat suspicious.

What I with my strength couldnt judge at all was how difficult exactly the decisions were between the ONLY move and the bad alternatives. Perhaps there is a forced development. I dont know. But I - holding for Hou - was deceived that White never failed. May others give their comment. It's already big feedback but nobody had the chess superiority to reason why possible exactly these moves could all have been played by a an average GM. That's too difficult to judge for me. Therefore I could only say that I couldnt believe it.

Look, if we all would always write honestly at the instant, then we had a much better discussion. Because you simply cant compare or even confrontate opinions out of the original moment and then longtime well thought analyses. But if we lower talented seekers would always refreign from telling what we thought this place would be so lively like the announcements of the ICGA journal which appears as a quarterly IMO.

Personally I find that Uri for one time more has shown how one should react. Trying to repeat the process and making a short conclusion. The rest will develop out of the following exchanges. But all this doesnt make invalid the original live comment as a possible truth. Which then could be refutated. This is what science is all about. We have a much too personalized debating tradition in this forum. At least I have the impression. Asling questions is often seen as a sign of mediocricy. In truth it's the beginning of everything.

[M ANSARI]

If she was text messaging in the toilet, then she should be disqualified even if she wasn't cheating. The rules are quite clear in that you cannot communicate with someone during a game. But in her case I might forgive her since unlike other chess playing girls, she is good to look at

It is quite easy to cheat if you have a mobile with someone on the other end giving you Rybka 3 output on strong hardware. On critical moves a good player like her could easily make very good use of that information. I am not saying she is cheating, but really no electronic devices should be allowed on any of the players during a game. If you are required to turn off your mobile devices on a plane why not in a chess match?

[Rolf]

Rybka does not play 28.Qg3 and prefers 28.g4(except fast time control)
and I used slow PIV.

This is correct. You found it. Now I dont remember if Kosteniuk played it quickly or not. I remember faintly that she played it although g4 was already as the absolute hammer move on the display. Next question, why I still wrote the message after move 36.Rxh7 and not so logically take the difference between g4 and Qg3 so seriously. I think I argued to myself that the played Qg3 isnt a losing move at all neither one with a negative evaluation. It's just not the absolute top move.

Since you showed this the question could be asked if a player of her strength who relied on outside help in positions where she had really to find the only possible positive move when all others wouldnt just lose directly but would have got a negative evaluation. But perhaps, Uri, it's already wrong to speak of ONLY moves, if it's only displayed as the only move with a minimal positive value.

I see, this could be the nature of my wrong judgement. I relied too much on sophisticated perhaps empty differences in the eval of a machine where a real GM possible wouldnt see a relevant difference at all.

Perhaps this is the crux of these top machines in chess. We, the majority of very weak players or experts as we like to call us as non-masters we get irrationally misled by the facticity of output numbers who could in real also be presented with different flowers. But there is no proven or reliable truth in such small differences of deci-numbers.

What do you think?

Still we could try to research if it's common that such non-superGM players have so many machine-approved moves in their games. Perhaps others now could become interested?

[Rolf]

If she was text messaging in the toilet... she is good to look at

Since I followed the broadcast on a server I had the moves and time control alone without seeing the toilet regions. But when I tought about the tech of the transmission I imagined how this could be done.

Kosteniuk has a magnet in her shoes. Say four in whole. And below under the stage there is the center of DoctorEvil with computers and hot coffee. Now the magnets are activated or deactivated. The day before I had another possibility in mind, this time for Hou. With her two [!!] radio transmitter antennas on her head! But unfortunately the KGB with the help of the of the NSA could re-direct the radio waves into magnetic waves so that Kosteniuk could ponder what the Chinese had cooked out.

Do you think I'm crazy?