Monday, July 23, 2007

Cheating incident in Poland

Cheating Incident In Poland

An incident of cheating occurred some days ago in the Polish town of Police, during the Tadeusz Gniot Memorial (11.07–19.07). One of the participants, Krzysztof Ejsmont from the host country, was expelled from the tournament after 7 rounds for the reason of "unsportive play".

20 year old non-titled Krzysztof Ejsmont was not considered a favorite of the tournament, with his Elo rating of only 2367. After a win in the first round against a much weaker player he continued his victorious series by 5 wins in a row against stronger opponents, among them were GMs M. Grabarczyk and V. Malaniuk.

The games themselves were excellent and full of sensational ideas. Suddenly it was discovered (independently by different participants) that many of Ejsmont's moves coincided with suggestions made by the program "Rybka", version 2.3.2. The detailed check-up of these 5 games showed that the percent of the coincidences was... 98%!

The full report was done on
Posted by Picasa


Anonymous said...

I hope he'll be banned for life.

Gambit said...

I'm not real aware of what the rules say about cheating but just a thought. I was just wondering, that until the guy doesn't get caught in the act (like that D.P Singh guy, I think, with the wireless device) then how do the regulating authorities accuse a candidate of cheating? Post each game, if this guys moves were coinciding with Rybka's lines up to 98%, although highly unlikely can we not say that maybe he had the same plans as Rybka. Now dont get me wrong, I know this sounds silly, but can you rule out that possibility, is what I'm asking? I'm not just claiming something silly, I just want to know what's another way of looking at this? Please share some thoughts readers.

Anonymous said...

It's sad!

Why cheat in this beautiful game?? :((

Wheter you are rated 2300 or 1300 or 1900 or doesn't matter!!! We can all enjoy the great game of Chess, can't we?

Cheating is something i would NEVER do!

Ban for life is indeed the way to go!

Anonymous said...

The link doesn't go to anything about this cheating...and cannot find it on that site either. Please post a working link, thank you

Anonymous said...

depends on positions. in a forced continuation any half decent player could get 100% coincidence with rybka. but 98% overall for 4 games is very unlikely. i would say definitely a case for the authorities at the tournament to confront the player. im not sure what exactly was this "confronting" at the polish tournament, but i think a straightforward immediate ban is not the right course of action. clarify the suspicion beyond any doubt and then, act according to the results.

Anonymous said...

First thing, we need to see the games.

Anonymous said...

It is impossible that 2360 player can achieve such result. 98% - Kasparov in best games achieved about 85 % - and not in 5 games in a row.

Anonymous said...

IIRC it was 98% because he made 'transmission' mistakes (played h4 instead of Nh4)

Games (quoting from pzszach forum):

Ejsmont K. - Grabarczyk B.
1:0, .

1. e4 e6 2. d4 d5 3. Sd2 Sc6 4. Sgf3 g6 5. c3 Gg7 6. Gd3 Sf6 7. e5
[7.e5 +0.52|d10 Rybka v2.3.2.w32]
7... Sd7 8. O-O
[8.O-O +0.45|d8 Rybka v2.3.2.w32]
8... f6 9. e:f6
[9.e:f6 +0.47|d7 Rybka v2.3.2.w32]
9... H:f6 10. b4
[10.b4 +0.54|d9 Rybka v2.3.2.w32]
10... a6 11. a4
[11.a4 +0.50|d8 Rybka v2.3.2.w32]
11... e5 12. Hb3
[12.Hb3 +0.87|d6 Rybka v2.3.2.w32]
12... e4 13. We1
[13.We1 +1.52|d6 Rybka v2.3.2.w32]
13... Se7 14. S:e4
[14.S:e4 +1.64|d6 Rybka v2.3.2.w32]
14... d:e4 15. W:e4
[15.W:e4 +1.72|d6 Rybka v2.3.2.w32]
15... Hf8 16. Ga3
[16.Ga3 +2.02|d8 Rybka v2.3.2.w32]
16... Sf6 17. Wee1
[17.Wee1 +2.50|d9 Rybka v2.3.2.w32]
17... Gf5 18. b5
[18.b5 +2.68|d7 Rybka v2.3.2.w32]
18... G:d3 19. G:e7
[19.G:e7 +3.76|d6 Rybka v2.3.2.w32]
19... H:e7 20. W:e7
[20.W:e7 +3.84|d6 Rybka v2.3.2.w32]
20... K:e7 21. Se5
[21.Se5 +4.05|d6 Rybka v2.3.2.w32]

Urban K. - Ejsmont K.
0:1, .

1. d4 Sf6 2. c4 e6 3. Sc3 Gb4 4. e3 c5 5. Gd3 d5 6. c:d5 S:d5 7. Sge2 Sc6 8. O-O c:d4 9. e:d4 O-O 10. S:d5 H:d5 11. Ge3 Se7 12. a3 Gd6 13. Sc3 Ha5 14. b4 Hd8 15. Hf3 Sf5
[15...Sf5 +0.34|d5 Rybka v2.3.2.w32]
16. G:f5 e:f5
[16...e:f5 +0.18|d6 Rybka v2.3.2.w32]
17. Gf4 a5
[17...a5 +0.21|d6 Rybka v2.3.2.w32]
18. b5 G:f4
[18...G:f4 +0.19|d11 Rybka v2.3.2.w32]
19. H:f4 Gd7
[19...Gd7 +0.18|d7 Rybka v2.3.2.w32]
20. Wab1 Wc8
[20...Wc8 +0.02|d6 Rybka v2.3.2.w32]
21. Sd5 We8
[21...We8 -0.22|d7 Rybka v2.3.2.w32]
22. Wfd1 Ge6
[22...Ge6 -0.22|d7 Rybka v2.3.2.w32]
23. Se3 g5
[23...g5 -0.29|d7 Rybka v2.3.2.w32]
24. Hf3 f4
[24...f4 -0.22|d6 Rybka v2.3.2.w32]
25. d5 Gd7
[25...Gd7 -0.40|d6 Rybka v2.3.2.w32]
26. Sf1 Gf5
[26...Gf5 -0.91|d6 Rybka v2.3.2.w32]
27. Wb2 Hd6
[27...Hd6 -0.94|d6 Rybka v2.3.2.w32]
28. Hh5 f6
[28...f6 -0.82|d6 Rybka v2.3.2.w32]
29. Hf3 Ge4
[29...Ge4 -1.10|d6 Rybka v2.3.2.w32]
30. Hb3 We5
[30...We5 -1.06|d5 Rybka v2.3.2.w32]
31. Wbd2 Kg7
[31...Kg7 -1.16|d11 Rybka v2.3.2.w32]
32. h4 h5
[32...h5 -1.25|d5 Rybka v2.3.2.w32]
33. Ha4 g:h4
[33...g:h4 -1.34|d6 Rybka v2.3.2.w32]
34. We1 Wce8
[34...Wce8 -1.10|d6 Rybka v2.3.2.w32]
35. H:a5 h3
[35...h3 -1.30|d6 Rybka v2.3.2.w32]
36. f3 Gd3
[36...Gd3 -2.60|d12 Rybka v2.3.2.w32]
37. W:e5 W:e5
[37...W:e5 -2.52|d9 Rybka v2.3.2.w32]
38. W:d3 Hc5
[38...Hc5 -2.73|d7 Rybka v2.3.2.w32]
39. Kh2 h:g2
[39...h:g2 -3.24|d6 Rybka v2.3.2.w32]
40. Se3 f:e3
[40...f:e3 -3.54|d7 Rybka v2.3.2.w32]
41. He1 Wg5
[41...Wg5 -11.81|d6 Rybka v2.3.2.w32]

Ejsmont K. - Malaniuk V.
1:0, .

1. e4 e5 2. Sf3 Sc6 3. Gb5 a6 4. Ga4 Sf6 5. O-O b5 6. Gb3 Gb7 7. d3 Ge7 8. We1 O-O 9. h3 d6 10. a3 Hd7 11. Sc3 Wae8 12. Ge3 Gd8 13. Se2grajac to posuniecie biale odrzucily propozycje remisu Se7 14. Sg3 c5 15. a4 Hc7 16. a:b5 a:b5 17. c4 b4 18. Ga4do tej pory jak wg partii Shirov - Milos 1999 Sd7 19. h4
[19.Sh4 +0.35|d11 Rybka v2.3.2.w32]
19... Kh8 20. h5
[20.h5 +0.57|d6 Rybka v2.3.2.w32]
20... h6 21. Gb5
[21.Gb5 +0.46|d13 Rybka v2.3.2.w32]
21... f5 22. S:f5
[22.S:f5 +0.66|d6 Rybka v2.3.2.w32]
22... S:f5 23. e:f5
[23.e:f5 +0.66|d8 Rybka v2.3.2.w32]
23... W:f5 24. Wa7
[24.Wa7 +0.48|d6 Rybka v2.3.2.w32]
24... Hc8 25. Sd2
[25.Sd2 +0.79|d6 Rybka v2.3.2.w32]
25... Gb6 26. Se4
[26.Se4 +1.75|d7 Rybka v2.3.2.w32]

Grabarczyk M. - Ejsmont K.
0:1, .

1. d4 Sf6 2. c4 e6 3. g3 d5 4. Gg2 d:c4 5. Ha4 Sbd7 6. H:c4 c5 7. Sf3 a6 8. Hc2 b6 9. O-O Gb7 10. Sc3 G:f3
[10...G:f3 +0.33|d9 Rybka v2.3.2.w32]
11. G:f3 c:d4
[11...c:d4 +0.34|d9 Rybka v2.3.2.w32]
12. G:a8 H:a8
[12...H:a8 +0.13|d8 Rybka v2.3.2.w32]
13. Sb1 h5
[13...Gc5 +0.05|d11 Rybka v2.3.2.w32]
14. h4 Gc5
[14...Gc5 +0.14|d9 Rybka v2.3.2.w32]
15. Sd2 Se5
[15...Se5 +0.00|d9 Rybka v2.3.2.w32]
16. b4 G:b4
[16...G:b4 -0.33|d9 Rybka v2.3.2.w32]
17. Sf3 d3
[17...d3 -2.12|d10 Rybka v2.3.2.w32]
18. Ha4 b5
[18...b5 -1.87|d10 Rybka v2.3.2.w32]
19. H:b4 d:e2
[19...d:e2 -1.90|d9 Rybka v2.3.2.w32]
20. Sd2 e:f1=H
[20...e:f1=H -2.03|d9 Rybka v2.3.2.w32]
21. K:f1 Hh1
[21...Hh1 -2.11|d11 Rybka v2.3.2.w32]
22. Ke2 Hd5
[22...Hd5 -1.98|d12 Rybka v2.3.2.w32]
23. Gb2 Sd3
[23...Sd3 -2.49|d11 Rybka v2.3.2.w32]
24. Hc3 S:b2
[24...S:b2 -2.57|d11 Rybka v2.3.2.w32]
25. H:b2 O-O
[25...O-O -2.62|d11 Rybka v2.3.2.w32]
26. Wc1 Wd8
[26...Wd8 -2.74|d11 Rybka v2.3.2.w32]
27. Hc2 b4
[27...b4 -2.74|d9 Rybka v2.3.2.w32]
28. Wd1 He5
[28...He5 -4.82|d12 Rybka v2.3.2.w32]

Ejsmont K. - Chojnacki K.
1:0, .

1. e4 e5 2. Sf3 Sc6 3. Gb5 a6 4. Ga4 Sf6 5. O-O Ge7 6. We1 b5 7. Gb3 d6 8. c3 O-O 9. h3 Sb8 10. d4 Sbd7 11. Sbd2 Gb7 12. Gc2 We8 13. a4 Gf8 14. Gd3 c6 15. Sf1 d5 16. Gg5 h6 17. G:f6
[17.G:f6 +0.16|d6 Rybka v2.3.2.w32]
17... S:f6 18. e:d5
[18.e:d5 +0.23|d7 Rybka v2.3.2.w32]
18... e4 19. d:c6
[19.d:c6 +0.43|d9 Rybka v2.3.2.w32]
19... G:c6 20. Se5
[20.Se5 +0.53|d10 Rybka v2.3.2.w32]
20... Hb6 21. Gc2
[21.Gc2 +0.74|d12 Rybka v2.3.2.w32]
21... Wad8 22. S:c6
[22.S:c6 +1.06|d11 Rybka v2.3.2.w32]
22... H:c6 23. a:b5
[23.a:b5 +1.10|d11 Rybka v2.3.2.w32]
23... a:b5 24. Sg3
[24.Sg3 +1.30|d13 Rybka v2.3.2.w32]
24... Gd6 25. Sh5
[25.Sh5 +1.31|d11 Rybka v2.3.2.w32]
25... S:h5 26. H:h5
[26.H:h5 +1.31|d11 Rybka v2.3.2.w32]
26... Gb8 27. Hg4
[27.Hg4 +1.43|d11 Rybka v2.3.2.w32]
27... Hc7 28. G:e4
[28.G:e4 +1.41|d10 Rybka v2.3.2.w32]
28... b4 29. Hf3
[29.Hf3 +1.44|d10 Rybka v2.3.2.w32]
29... Hh2 30. Kf1
[30.Kf1 +1.49|d10 Rybka v2.3.2.w32]
30... We6 31. Wa8
[31.Wa8 +1.55|d9 Rybka v2.3.2.w32]
31... Wf6 32. He3
[32.He3 +1.60|d9 Rybka v2.3.2.w32]
32... b:c3 33. b:c3
[33.b:c3 +1.62|d10 Rybka v2.3.2.w32]
33... We8 34. Wb1
[34.Wb1 +1.87|d10 Rybka v2.3.2.w32]
34... Wfe6 35. f3
[35.f3 +1.76|d11 Rybka v2.3.2.w32]
35... Gf4 36. W:e8
[36.W:e8 +1.79|d12 Rybka v2.3.2.w32]
36... W:e8 37. Hg1
[37.Hg1 +1.85|d12 Rybka v2.3.2.w32]
37... Hg3 38. Hf2
[38.Hf2 +1.74|d10 Rybka v2.3.2.w32]
38... Hh2 39. Wa1
[39.Wa1 +1.09|d15 Rybka v2.3.2.w32]
39... f5 40. G:f5
[40.G:f5 +1.91|d9 Rybka v2.3.2.w32]
40... Ge3 41. Ha2
[41.Ha2 +2.06|d11 Rybka v2.3.2.w32]
41... Kf8 42. Ha3
[42.Ha3 +2.29|d11 Rybka v2.3.2.w32]
42... Kg8 43. Hb3
[43.Hb3 +1.16|d9 Rybka v2.3.2.w32]
43... Kh8 44. He6
[44.He6 +2.27|d13 Rybka v2.3.2.w32]
44... Hg1 45. Ke2
[45.Ke2 +2.26|d14 Rybka v2.3.2.w32]
45... H:g2 46. Kd3
[46.Kd3 +2.25|d15 Rybka v2.3.2.w32]
46... Hd2 47. Kc4
[47.Kc4 +2.34|d15 Rybka v2.3.2.w32]
47... Wd8 48. Hg6
[48.Hg6 +4.79|d10 Rybka v2.3.2.w32]
48... W:d4 49. Kb3
[49.Kb3 +12.01|d7 Rybka v2.3.2.w32]
49... Wb4 50. K:b4
[50.K:b4 +299.90|d12 Rybka v2.3.2.w32]
50... Hb2 51. Kc4
[51.Kc4 +299.95|d8 Rybka v2.3.2.w32]

Anonymous said...

Gambit - Agree with your comments. 98% is statistically almost impossible and while it is almost certain that this indicates cheating, it is not proof. What would happen if the expelled player decided to file a lawsuit (particularly if he was a professional).

What amazes me is that not only do certain players cheat but they do so very stupidly - beating everyone in sight with fantastic ideas. You think that they would at least learn to do it in a more circumspect manner e.g. just using the method for critical positions in critical games.

To your point on proof I think it is essential in each case to find out exactly how it is being done (e.g. wireless devise, hand signals etc)otherwise you have no defense to someone who does it properly in future.

The question arises also as to what would happen if a 2700+ player made 98% the same moves as Rykba? Could you throw them out of the tournament without finding the actual method of cheating, since it is conceivable that in certain games they could easily play the same moves as Rykba. e.g. One US player (Silman?, Seirwan?)did an analysis of games from old world championships and found a lot of moves (90%+?) that were first or second choices of Fritz9.

tfk said...

To anon, you made the comment "to someone who does it properly in the future". That is an oxymoron! You cant cheat properly!

Anonymous said...

The organizer of the tournament Mieczyslaw Manik says that before the 8th round Ejsmont admitted to using Rybka in games 2-5 but 20 minutes later he denied cheating.

Anonymous said...

i mean in games 2-6, not 2-5

Anonymous said...

Do the tournament directors have on-site access to Rybka, Fritz, and the other top programs (i.e. shredder, Junior, etc)? I mean, if they suspect cheating can they do an "on the spot" analysis?

If cheating can be proven, and this is quite important as it must be proven, then the line must be drawn and cheaters banned from tournament play.

I know of no other game (i.e. Go, checkers, backgammon) in which players are being accused of outright cheating as is now the case in chess.

It doesn't matter what rating you are or whether you are the WCC. If, and only if, you admit your guilt or the method of cheating can be proven, then repercussions/punishment should be swift and permanent.

If a titled player is caught cheating, I believe they should be banned for life and have their titles (be they NM, IM, or GM) stripped from them.

Seem harsh?

It must end.

Finally, if TD's have access to such software during the games, it is always possible that they could be "paid off" to assist a player.

Rules should be in place for them and thorough background checks done.

You never know who might be assisting a player...even a TD. I don't think it's happened so far but people should be aware of the possibility.

Jerzy said...

The circumstances of the incident include a delegation of several tournament players to the accused of cheating and the delegation promised not to make it public in case the suspect would have withdrawn from the tournament.

Furthermore the evidence should be checked by an independent jury d`appel and if confirmed it is quite substantial like leaving one`s fingerprints on a crime scene.

It would be advisable to find out what were the means of communication between golden fish and the suspect and to find out who supported the suspect in the cheating.

So it is a long way to Tiperary to make all things clear.

Anonymous said...

Suddenly it was discovered? At a tournament like this, they should have someone going over the games of all unusual upsets routinely.

Anonymous said...

Anonymous said...
Cheating is something i would NEVER do!

And your name was?

Jerzy said...

"At a tournament like this, they should have someone going over the games of all unusual upsets routinely."

Oh, it was only a local chess tournament in which only a couple of titled players had taken part. And one of them was outplayed in 15 minutes by the suspect so he started to doubt his honesty. However I advise to check his proofs of cheating before the final judgement.

BTW It is much easier to cheat by a titled player in a tournament where there are many such players like him. No one routinely checkes the coincidence of moves made by him with moves made by chess programs because of mutual confidence unless it is a WCC match between Topalov and Kramnik :-)

Jon said...

Funnily enough, the tournament was eventually won by someone else with a comparatively low rating!

fh said...

98% coincidence rate between a 2300 player and a 3100 program is evidence certain enough that any court would uphold it.

It's like publishing your own version of Harry Potter only changing the names and when someone accuses you of plagiarism claiming that it just so happened randomly that you wrote the same book without knowing about Harry Potter!

Of course finding the channel of communication is crucial to prevent more intelligent people from cheating more intelligently (proper cheating = cheating where you don't get caught) in the future.

Jerzy said...

"It's like publishing your own version of Harry Potter only changing the names"

However there are no procedures so far as in finding out plagiarism or in discovering fingerprints in a crime scene. FIDE and other chess organizations should take appropriate measures in counteracting the cheating with the means of chess programs.

Anonymous said...

With so many chess fans also having an interest in computers and mathematics it should not be difficult to find someone who could do a simple experiment - at some large open (World, National, Isle of Man, etc) - just get the games from the bulletin for a statistical sample of players 1800, 1900, 2000, 2200, 2400, >2400 and compare their play with the computer - this will help give an idea of the range of times strong players 'coincide' with computers - as others have noted, very strong players in forcing lines will of course play just like the computer.

?thesis topic for someone?

Anonymous said...

I do not know about this Krzysztof Ejsmont fellow's actions or possible confession, but in general...

Nobody should be fully accussed of cheating on the basis of their high quality moves alone, even in extreme cases.

(Else every low rated player who has ever entered an open event deserves to have his entry fee ante refunded.)

Warn him, sure. Tell him you are suspicious or even doubt him. Search his hat and clothes etc, sure.
But without a smoking gun, his victories must be tolerated.

Even with today's technologies, an intense search and investigation can find the smoking gun if there is one.

Low tech hand or body signals from a co-conspirator spectator may be the wisest mode of cheating.
The TD should need no hard evidence to eject any spectator. In am imperfect world, ejecting a few spectators may be the least of evils.


Anonymous said...

it is crazy situation.
That case the person was not going to the bathroom and has been judged on the base of the analysis done by conucrrence. Nobody has found any prove except computer analysis. It is very doubtful proove because it can be manipulated. The move selected by program depends on time given to think. It is very easy to stop calculation when the program selected desired move.
It is dangerous now to play good not being known player. Anyone could tell that your play correlates with Rybka, Hiarcs, Shredder, Nimzo etc. It is easy to find 'proper' programm and made accusation. My play is correlating with Shredder i.e.:-)


Tim Harris said...

It is my opinion that there must be a "smoking gun" found or else, in legal terms, there is room for "reasonable doubt."

Now, I concur that it is highly improbable (almost impossible that this was a mere coincidence) that a 2300 player can match Rybka's moves with 98% accuracy.

However, if the person was not acting in a dubious manner, such as leaving the area, in addition to not having any devices found upon his person after a search, and the only thing that raised concern was his defeating players much better than him, then it cannot be proven beyond a reasonable doubt that he was cheating with Rybka.

Now, don't get me wrong. There must have been a way he was getting assistance it just was not found nor explained in any other way than analyzing his games with Rybka and seeing the matching moves.

On this basis alone, without a smoking gun, it ultimately seems to come down to an allegation with no empircal evidence to support it. That, in and of itself is nothing. You must have solid proof.

If no devices were found and no strange behavior was obvious, I cannot see how he can be punished even though as chess fans most would agree he had to have assistance.

This is a strange case. If no devices were found, I wonder how he obtained assistance from Rybka?

There must be an explanation because as Sherlock Holmes character said, "When you rule out the impossible, whatever is left, no matter how improbable...must be the truth."

So, once the impossible is ruled out, what, I ask, is left?

I look forward to everyone's opinion.

Thank you.

Tim Harris

Anonymous said...

There is no information here to make a judgement. We need solid information or else we are simply speculting about someone else.

I can only assume that more information will be made available in the near future. Interesting story.

Anonymous said...

Chess is giving way to fishes like Rybka. It's only a matter of time when Fritz starts cheating, using Rybka, too? Or vice-verse?

Vinay said...

Its sad that such a highly rated player cheats. There is no hope for lower rated tournaments!

- Vinay

Anonymous said...

He probably used an inner ear canal receiver that is invisible without examining the inner ear with an ear probe (I linked to the vendor's site for one of these devices some months ago in this forum).


He uses the new eyebrow transmission technique originally proposed by the KGB and then later perfected by one of my associates in Interpol. No wires, no receiver, it is virtually impossible to detect, even when you know what you're looking for.

Anonymous said...

With the current technology and micro-technology progress, it will be soon impossible to find devices. How do you expect to find communication device which one has implanted in his body, for example?

98% coincidence (note that remaining 2% are mistakes in communication like playing Nh4 instead of h4) is proving the case. People are sentenced to jail using less obvious traces of the crime.

According to the participants, Ejsmont behaved strangely:

- he was sitting behind the board without leaving it at all, while during old tournaments he used to walk a lot (the tournament was not relayed, people suspect he was wearing some camera)

- he played very fast (in the game in which he crushed GM Grabarczyk he used just 15 minutes of his thinking time for the whole game)

It was also impossible to search him, for two reasons:

- he started to be suspected only after 6th round, and knew that 'something is happening', so during 7th round he was probably clean (he agreed to a short draw then, in spite of he played lower level opp and earlier crushed GMs)

- the law in Poland restricts activities like that to the police only, and it is very unlikely police would find it reasonable in this case

Anonymous said...

Circumstantial evidence can certainly be enough to 'convict' someone. You don't need transmitters and confessions. This guy is guilty as charged ...

Anonymous said...

Level of proof depends on level of punishment (e.g. balance of probabilities in non-criminal court cases). You don't plan to execute this guy (I assume) so a smoking gun is not necessary to ban him. The posting of the games 6:16 makes it clear that he played exactly the move of Rybka every time (not one of Rybka's first two suggestions 90% of the time - a very different scenario) except for when he misheard the move. So not only was he using Rybka, he was either being very foolish or happy to get caught. Maybe he was trying to make some sort of point. Unless that is the case, the evidence is strong enough for an authority overseeing fair play in a sport to ban him, even if not enough to convict him in a court of law.
PS Never been banned from a golf club on hearsay evidence from your girlfriend's husband?

Anonymous said...

Look at the games!
I'd prefer to believe a guy who said someone had planted an electronic device in his cap before anyone who said that he played EVERY MOVE exactly the same as Rybka's FIRST choice (not 90% of first or second).
You're not going to execute him so "smoking gun" of criminal proceedings not required. You can ban him.
Except...he so clearly was happy to get caught, or would have varied things slightly. Is he trying to make a point?

Anonymous said...

On July 25, 2007 03:07:00, Anonymous writes: "So not only was he using Rybka, he was either being very foolish or happy to get caught. Maybe he was trying to make some sort of point."

Interesting, perhaps he is a small fish in some larger scheme and he was merely testing the waters or limits of what one could get away with. What if there are more, less obvious (like only using the computer at critical moments, or when cued for help) and higher rated players in on this scheme as well? This may bring some of our highest rated GMs' play under scrutiny .

KWRegan said...

Quick reactions and a request:

(1) ChessToday seems to be an e-mail newsletter, so I'm not sure if even signing up for the Free Trial would include that report.

(2) To answer Tim Harris, my anti-cheating website's page testing the 2006 World Open allegations against Eugen Varshavsky as positive has this statement toward the bottom: "Current working consensus in the chess world, reflected by National Director Steve Immitt at the end of (the move-match section of) the March 2007 Chess Life article, is that match-rate statistics must be accompanied by some other primary evidence---such as physical or eyewitness evidence. This site supports this policy." (The translation below seems to indicate by its speculations that no physical evidence was found.)

The rest of my paragraph has a note to the effect that if you look thru enough of your games for statistical evidence your opponent was cheating, you'll find it. However, having *four* games in the same event avoids this caveat, and if my tests confirm the findings, they will be statistically significant---this answers some other commenters here. If there is no physical or eyewitness evidence, then this is a difficult test case of whether stats alone can be practically conclusive...

(3) I found a Polish news story on this at
I've quick-translated it by cutting and pasting chunks of text into, fixing just a few things by my general knowledge of Slavic languages:

Moves By Computer
This year's memorial of Tadeus Gniota will be remembered by moves from a computer by means of an amazing scandal. 20-year-old Krzysztof Ejsmont was the revelation of competition from hetman 20 chenstochowa about. However, after seventh round - as yesterday we wrote in (to) already " voice " he has been excluded from the competition -. Why?

Money-winning champion

It played with number 15 Ejsmont Policach. It means, that it was in bunch in competition chess ranking fifteenth 35 ( so called ) starting ELO szachistów. Therefore, it was not included for bunch of favorite. So, that was surprise, when young sportsman won party behind party - and it with considerably better rivals. There is national (local) champion only Ejsmont, however, it overcame international champions and absolute master. Suspicion has emerged, that sportsman does not play < self-service store > - that behind assistance of transmitter and it contacts with (from) somebody receiver, who sits at strong chess computer. And it prompts computer fairest movement (traffic) Ejsmontowi.


The scandal started on Sunday, when the fifth round was played. Then, it has overcome absolute master of miroslaw with (from) playing black mazovian Ejsmont Grabarczyka Polfy Grodzisk. It sat after defeat over chessboard long Grabarczyk. It thought how (as) it, that it playing with (from) before young person has demonstrated such ability. Something at the end he (it) tknęło. Record of party begin analyzing.

It records moves in the course of all chess duels and among others, it is as (serve) for analyses opponent ( later ). It has discovered Grabarczyk, that Ejsmont's moves were consistent with a certain chess program called " Rybka ". The Polish husband of WIM Iweta Radziewicz is author of this program. Program is characterized outsized power of game.

Then, it has come for I - Grabarczyk main judge tells - competition polickiego Wenanty Błażejczak. To monday about hour 14.30 -, it has reported before sixth round me Grabarczyk, that play < game > " unethically " Ejsmont. Then, we have started analyzing of its (his) other party.

Rybka is stronger than Kasparov...

And five analyze from second (other) for sixth. As it happens, that simple (easy) is from completion of chess theory ( initial move, because it is possible to find it (them) play in each book about chess according to program party ) to the end " fish " Ejsmont. - I.e. it choose these moves, suggested which (who) in (to) first sequence " fish ", or as fairest it explains - Błażejczak. It single out moves from (with) said five parties 102 - Ejsmonta. Consistent was from it with (from) 100 " fish ". How is this program? Explain it on base of ranking best ELO. Plays at the level of near 3000 " fish ", but ranking has 2367 points presently Ejsmont. Says it a few? But they let's add it, that with (from) play fairest Policach szachistów, ukrainian has 2543 Wadim Szyszkin ELO. But does not have present champion of world in chess 2800 ELO! - Garri Kasparow As it played, it had over 2800 a bit chief has informed for affairs sports - competition polickiego, rafal Szyszyło ...

Transmitter in Pen?

Does this make Ejsmont a genius?

It is not play according to I person in state (condition) at the level of 3000 - practically ELO, maximum human reliability has has emerged it over 2800 a few it thinks (consider) - Błażejczak. So, bright (plain) has become for play Policach szachistów, that somehow must be jointed with program " fish " Ejsmont. Transmitter must have - and receiver organizer explains - competition polickiego, mieczyslaw Manik. It contacted with (from) somebody - due to it, who sat at computer and it introduced for program of move from party Ejsmonta. Where implicit transmitter had receiver and? It is solutions great deal. From small in ear (ear) kulki, by special plate under sheet of paper (rationing coupon) with record, for watch sprzężonego, if (or) pen. So, suspicion has has emerged about dishonesty of sportsman support circumstantial evidences. What and further? I have called small group - szachistów story continues - Błażejczak. It was compound (composite) < lodge > from two absolute master -, two (two) international champions and playing competition Pawła Grafa. It doctor, science officer, great moral weight. And we have chosen following (step) solution mr. Graf, which (who) was not interested one of farthest place in competition along with with (from) victory ( occupy (deal with) ) talk Wadimem Szyszkinem Ejsmontem and match will suggest it (him) some. Two solutions were - it explains - Szyszyło. - Either (or) it will admit Ejsmont, it will pronounce judge and it will recede from profession (contest), either (or) it will not admit , competition will finish, it will take award for victory ( 4 thousand zloty - dop. Roadsteads. ) But everything we will report polish chess union < relationship (association) > and international organization FIDE. First of all, however, we wanted to see reaction of defendant too.

He did not reflect long

It has pronounced suspicion Graf Ejsmontowi. It was not surprised defendant according to its (his) rate about cheating young szachista and it has admitted . But condition has put - that it will inform nobody. Then, it has beaten off Graf, that it plays impossible to the end, but then about knowledge < know > it szachiści, which (who) with (from) before they played. It has pronounced in spite of it Ejsmont, that it agrees and it will enter by eighth for judge round. However, it has asked about break, in order to cool down. --minute phone has executed these breaks in time 15 - near it relates - further Szyszyło. It has appeared on hall - after it and. All have been ousted . It said, that they be jealous some it (him), that competition of life plays , etc. however, could not convict these arguments. It, which (who) with (from) before play return e.g. note, that it played in complicated (elaborate) positions fastly very, where reflection was much needed. Bogdan has counted Grabarczyk, that it needed in most important period of party on reflection hour and 40 minutes. But over 15 minutes a few Ejsmont.

I take the last part as saying that Ejsmont moved quickly. I may be able to tell *how* quickly if there are cases where Rybka changes to another move at greater ply depth.

Does anyone know how the testing was done? Are there any checkable log files? The scientific and moral due-process issues here should be clear, especially when people are being named in public...

Anonymous said...

Very said (IMO) end of this story.

PZSzach (Polish Chess Federation) just decided not to punish Krzysztof Ejsmont in any way.

The ruling (in Polish) is here:

The message is rather short and lack deeper analysis. In general: they decided that even incredibly high coincidence with Rybka is not a proof.

Considering organizers are not allowed to search the player, and police does not consider those problems to be of their interest, feel free to cheat in case you play a tournament in Poland.

KWRegan said...

Google's own translation is much better than the one I used before: here is a TinyURL link to the resulting translation of its posting on a Polish chess blog.

On the same blog, I found a very recent post (Dec. 28, 2008) of remarks by Elista 2008 Grand Prix tournament competitiors on, "Is control of computer fraud necessary for world championship tournaments?" I note in particular Dmitry Jakovenko's comment:

"I agree that this is a very important question, although the debate about computer fraud ucichła bit. I think that there should be such a way as to find and prove that the player uses a computer at the party."

The one word missed by Google means "has subsided", as one can see via this dictionary which has a really useful prefix-search feature, so you can type the first few letters and get a list of possible words that extend them. I don't know whether by "prove" he meant by examining the game---the only other kind of proof would be finding physical or behavioral evidence during the game, and/or getting accomplices to confess. But I can say I still agree with the comments by Tim Harris and others in this thread, and with the gist of the Polish Chess Federation's ruling, that evidence from the game's moves cannot be dispositive of guilt by itself. The statistical modeling needed to bring this to the level of standardly accepted uses of statistics as evidence in court cases is proving fiendishly difficult, and at least from my vantage, alas has a long way to go. At least my efforts are now directed to larger positive rather than negative consequences, such as an absolute foundation for the Elo rating system.

KWRegan said...

Oops, my second link went to the same place as the first, i.e. to Hinda Sledziewski's blog entry for the Polish CF ruling itself. On re-trying this, it seems Google keeps the original frame's URL, so one has to do a new search in a new window for the new article in Polish, then click "Translate this page" again. Using a TinyURL also provides a simple way to test before posting, so the corrected link is here.

Meanwhile I found much more about the Ejsmont case on Hinda Sledziewski's blog, most notably this account by Marcin Krysztofiak.

rick said...

Ban for life!