|
< Earlier Kibitzing · PAGE 17 OF 22 ·
Later Kibitzing> |
| Sep-20-10 | | dpruess: littleRiver is not the owner of chess.com. |
|
| Sep-20-10 | | whatthefat: <bartonlaos: These are some of the numbers that chess.com used to help make their determination - Percentage agreement with Deep Rybka 3:
Deep Rybka 3 x64 Hash:256 Time:30s Depth:12-20ply
AMD Phenom x 4 2.30Ghz 4GB DDR2 RAM
YelenaDembo (Games: 20)
Top 1 Match: 530/723 ( 73.3% )
Top 2 Match: 638/723 ( 88.2% )
Top 3 Match: 676/723 ( 93.5% )
Top 4 Match: 698/723 ( 96.5% )
Here are the numbers from a normal correspondence player: 9th Correspondence Chess World Championship 1977-80** 1st Place:
Oim, Tonu O. (EST) (Games: 14)
Top 1 Match: 263/433 ( 60.7% )
Top 2 Match: 352/433 ( 81.3% )
Top 3 Match: 388/433 ( 89.6% )
Top 4 Match: 400/433 ( 92.4% ) >
There are 2 very clear flaws in this method of comparison: 1) The games being analyzed have in a sense been cherry-picked. Analyze any player's matching rate with Rybka at the end of a hot streak and it will be elevated. How were the games of the world champions selected? 2) The matching rate will clearly depend on the strength of the opposition. Against relatively weak opponents the matching rate will likely be substantially higher. It is thus totally invalid to compare her matching rate against players rated ~2200-2600, to the matching rates of correspondence world champions playing against the absolute elite. |
|
| Sep-20-10 | | wwwchessdev: "littleRiver" is NOT the owner of Chess.com. I have no idea who the imposter is, and I'm sorry they have completely misrepresented Chess.com. I am the owner of Chess.com (see my status: http://www.chess.com/members/view/e...). I have no comment about Yelena other than that her account on Chess.com is no longer active and I wish her the best in her continued chess career. |
|
| Sep-20-10 | | wwwchessdev: I also urge those who are posting on this topic to stop attacking each other to exercise prudence and respect. I'm sad to see this blown out of proportion. |
|
| Sep-20-10 | | Jim Bartle: OK, wwwchessdev, if you are in fact owner of Chess.com, what recourse did Yelena Dembo have to challenge chess.com's determination that she had been cheating? Any way to appeal, defend herself, state her case? Just curious. |
|
| Sep-20-10 | | darthpanda: Just to clear something up--the top 3/4 match-up methodology is something that was put forward by (but not created by) a respected member of the community devoted to eliminating cheating on chess.com. Its purpose is to help members determine when enough evidence of engine assistance exists to report a player to the staff for further analysis. A great deal of the criticism towards this methodology not being valid was by members who were later banned for violating the TOS regarding engine assistance in games. (AKA "Cheating") Chess.com may or may not use that method of move analysis, but even if they do, there are other steps they use to determine if someone is cheating. They don't mention what they are, but consider this: They have IM's and GM's on their payroll to provide content. How easy is it to throw a few of them some extra money to use their knowledge and expertise to determine if someone is making "human" moves or not in a situation like this? Given a large number of games, with enough non-database moves per game to give an appropriate sample of moves to analyze, there was enough consistency between her play and that of chess engines to exceed chess.com's standards for what is considered to be receiving engine assistance. That doesn't mean she's a horrible person or hasn't earned the respect due her title in the chess world. @Jim Bartle--Erik (wwwchessdev) is in fact the owner of chess.com. The recourse given to people who's account has been closed for cheating is simple. On the "homepage" for their account, there is a link they can click to contact support so they can appeal chess.com's closing of the account. It says something to the effect of "This account was closed for cheating. If you feel this was in error, click here." It's pretty hard to miss. As for the specifics of what happens from there, that would be something Erik or a chess.com staff member would have to answer, but as a member, can tell you that starts the process of stating her case to appeal the decision. |
|
| Sep-20-10 | | chesswizz1964: *bartonlaos*
I think Yelena plays a beautiful game, but to call Tonu Oim a 'normal correspondence player' is little bit unfair!
Tonu Oim is correspondence GM and first to win ICCF World Championship in correspondence chess twice, in 1983 and 1999.
http://en.wikipedia.org/wiki/T%C3%B...
Very good correspondence chess player. One of the best ever. |
|
| Sep-20-10 | | Jim Bartle: wwwchessdev, apparently the proprietor of chess.com, seems to have come and gone, leaving this message: "I also urge those who are posting on this topic to stop attacking each other to exercise prudence and respect. I'm sad to see this blown out of proportion." Rough translation: "Keep moving, keep moving. Nothing to see here. Keep moving." |
|
| Sep-20-10 | | Jim Bartle: Thanks, darthpanda. That pretty well answers my question. |
|
| Sep-20-10 | | atudor: I can assure everybody here that littleriver is not the owner of Chess dot com or any part of the staff.I have been a member of Chess dot com for 2 yrs. The site is one of the best on the web if not the best. The owner and staff have very high morals and would never do anything with out fully investigating the facts. I understand this is a bad situation but believe me Chess dot com has very high standards and would not do anything like this without cause. I'm personally very Sorry this has happened. |
|
| Sep-21-10 | | Fezzik: The entire chess community, and especially chess.com, will benefit if the chess.com shows complete transparency in how it determines engine abuse. Chess.com's initial statement was that it was "100% certain" that Dembo cheated. Such a statement belies a lack of statistical rigor, at least by the person making the statement. If there is an excellent method for catching cheats, the rest of the world chess community would benefit from seeing it. More importantly for chess.com, it would benefit by proving that the website was careful and thorough before banning yet another professional chess player. |
|
| Sep-21-10 | | CCplayer: I especially like the statement made by dpreuss that they have not yet banned anyone innocent. How does he know?? Did they all confess in writing to get it off their chest??
That's one of those things that sounds good, but has no value whatsoever. I'm inclined to side with Fezzik in his last post. Unless the process for detecting engine usage is made public and available for people trained in statistics to critizise and make comments on it has no value whatsoever. Then it is just an opinion of chess.com that you can choose to believe or distrust. I also agree that the statement of 100% certainty is by itself ridiculous - there are only probabilities here. Being an outsider I have since yesterday heard from a friend with an account on chess.com that there's a pretty large mob of shouters who seem to think that anyone who achieves a high rating on the site is automatically using an engine. He (my friend) also states that chess.com staff has a very weak attitude towards this and more or less are considered to be in silent agreement. If your reward for winning more games than the average player on chess.com is to get your name dragged in the mud in public, membership seems like a dopubtful proposition...... My guess is that this will lead to all strong players leaving the site. |
|
| Sep-21-10 | | newbieatchess: in an early comment someone said why the analysis results didn't match her otb games, well cc has long time controls and don't think its fair to compare the two its like comparing a game played at 2hr/40 move to a bullet game. |
|
| Sep-21-10 | | chesswizz1964: Ok so why not we see some games where Yelena plays this perfect program chess?
No games, no proof.
Yelena innocent Greek chess Goddess!! |
|
| Sep-21-10 | | bullyboy12708: You want games that show her matching Rybka yes? ok i'll get you them games them when you analyse them for yourself you will see exactly what we saw! The reason chess.com do not publish their methods is because once all the cheaters see it they will adapt to beat the system. You see it all the time with anti crime measures. A new system comes out and within days/ weeks the criminals beat it because they learn to adapt. |
|
| Sep-21-10 | | CCplayer: Anti-crime measures? You have a great sense of perspective Bullyboy. |
|
| Sep-21-10 | | bullyboy12708: Oh come on you know the score if cheaters find out what they face they will adapt of that i'm certain |
|
| Sep-21-10 | | chesswizz1964: See?
Zero proof. |
|
| Sep-21-10 | | CCplayer: It was the word crime that caught my attention. That usually means an act punishable by criminal law. I doubt that chess.com has that authority.
Regardless I think your attitude is wrong. It is never better to hide the detection-methods since then it will always be open for debate whether the "verdict" is fair or not. Those who are so desperate to use an engine in CC play that they are willing to work hard on it will always be able to do that regardless of any countermeasures. Just looking at the evaluation of resulting positions and getting information on the opponents strongest will give a huge advantage. |
|
| Sep-21-10 | | bartonlaos: chesswizz1964, I didn't mean any disrespect to Tonu Oim by calling him a 'normal' correspondence player. What I meant to distinguish was players accused of engine-like play, and players known not to have used computers at all (normal controls). I used quotes around the term 'normal' in a subsequent post. I've been taking some time to review Yelena Dembo's games through the Wayback Machine. What I've found is that she is the higher rated player in all the Chess.com games that I looked at. Not by a little, but by often hundreds of points. When such a difference exists in ratings, it's easy to follow the winning path, and have a higher-relative agreement. The large relative rating differences also helps to explain her near-flawless record. Anyway, I've found a source for the twenty games that were analyzed and will post shortly. The games spread out over 2 years, for some reason the source skipped 2009. Her rating in 2008 at 2450 was no where near peak of about 2700. The 2700 rating for cc at chess.com is within the limits for someone who has a FIDE rating of about 2500. The point is that she was a much stronger player in 2008 relative to her ccChess.com rating, and accentuates the real differences in ability between her and her opponents, should give her a higher percent agreement. Nine of the 20 games used to accuse her were taken from 2008, and these opponents averaged a Chess.com rating of 2250, so there existed an extremely large relative difference in strength for those games since her FIDE rating was 2450 (some opponents pushed a 500 point relative strength difference = subtract about 100 Chess.com points to get USCF rating, and subtract about 100 points USCF to get FIDE rating). Messy, sorry. Overall, there was about a 300 point rating difference over 19 games of her 20 games, with an 8 point difference in the most recent game. She never played a stronger opponent in the batch that was tested. Here you may find the games:
http://lousyatchess.blogspot.com/
The serious flaw is a lack of proper normal controls. I've tried to correct this in my next post. |
|
| Sep-21-10 | | bartonlaos: Here is my effort...I've used a player called Slow_Burn to help equate Normal players from Yelena and other accused "cheaters": ...........n=3 Slow_Burn vs. Normal.............
Slow_Burn (2700) vs. Normal Player-1 (2700)
http://www.chess.com/echess/game.ht...
Slow_Burn (2750) vs. Normal Player-2 (2750)
http://www.chess.com/echess/game.ht...
Slow_Burn (2800) vs. Normal Player-3 (2750)
http://www.chess.com/echess/game.ht...
...........n=3 Slow_Burn vs. "Cheater"...........
"Cheater" 1 vs. Slow_Burn (2800)
http://www.chess.com/echess/game.ht...
"Cheater" 2 vs. Slow_Burn (2800)
http://www.chess.com/echess/game.ht...
"Cheater" 3 vs. Slow_Burn (2800)
http://www.chess.com/echess/game.ht...
................Yelana's Game...............
Slow_Burn vs. Yelena Dembo
Yelena Dembo vs. Slow_Burn (2800)
http://www.chess.com/echess/game.ht...
................Yelena's Game...............
There isn't a Normal opponent for Yelena other than Slow_Burn, since the ratings don't pair up with hers. So I took one of the Accused Cheaters who played against Slow-Burn, and matched them with a Normal player and also another 'Cheater', to help try to understand agreements, but I haven't run the numbers on any of these: ........"Cheater 3" (Nosferatu) vs Normal and Cheater.... "Cheater" 3 vs. Normal Player-3 (2750)
http://www.chess.com/echess/game.ht...
"Cheater" 3 vs "Cheater" 4 (2800)
http://www.chess.com/echess/game.ht...
......... ............. ........
Granted it's a relatively small population to choose from, but that's all the time I have, lol! By the way, run the numbers on this game, it didn't qualify because the FM was under 2200, but I think it should have because of relative strength: Yelena Dembo vs FM_Eric_Schiller
[Event "Titled Player Championship - Round 1"]
[Site "Chess.com"]
[Date "2008.03.01"]
[White "FM_Eric_Schiller"]
[Black "YelenaDembo"]
[Result "0-1"]
[WhiteElo "1827"]
[BlackElo "2464"]
[TimeControl "1 in 3 days"]
[Termination "YelenaDembo won by resignation"]
1.d4 Nf6 2.Nf3 g6 3.c4 Bg7 4.Nc3 O-O 5.e3 d6 6.Be2 Nbd7 7.b4 e5 8.O-O a5 9.b5 Re8 10.a4 Nb6
11.Qb3 exd4 12.exd4 c5 13.bxc6 bxc6 14.Bf4 Be6 15.Rab1 Rb8 16.Qa2 Nh5 17.Be3 Qc7 18.Rfe1 d5 19.Ng5 Bf5 20.c5 Bxb1
21.cxb6 Qxh2+ 22.Kxh2 Bxa2 23.Nxa2 Rxb6 24.Bxh5 gxh5 25.Nf3 Rb2 26.Nc3 Bf8 27.Rc1 Reb8 28.Ne5 Bd6 29.f4 Bxe5 30.fxe5 R8b4
31.Rf1 h6 32.Rf6 R4b3 33.Rxc6 Rc2 34.Nxd5 Rxc6 35.Ne7+ Kf8 36.Nxc6 Rxe3 37.Nxa5 Ra3 38.Nc6 Rxa4 39.d5 Ke8 40.Kg3 Rc4
41.Kf3 h4 42.Ke3 Rc3+ 43.Kf2 Rc5 0-1 |
|
| Sep-21-10 | | bartonlaos: One thing, if you run the numbers on the 2nd part, and just want to look at the Slow_Burn vs Yelena game, then keep track of Slow_Burn's agreement as well, since they are a normal player with the same rating as Yelena. |
|
| Sep-21-10 | | CCplayer: Are the games used as proof cherrypicked?? Then it loses all credibility! |
|
| Sep-21-10 | | cotoi: CCplayer: chess.com does not disclose their cheating detection methodology and nobody except the staff has any idea about the criteria they use. All we can do is take some games Yelena Dembo played there (if we had) and compare them against an engine. I wouldn't be so concerned if the games are cherry-picked or not. If you have a relatively strong player who is cheating, would you expect him to use his engine even when playing against a 1200 beginner? |
|
Sep-21-10
 | | tamar: This isn't the Tour de France. You can't expect chess.com to convene a special commission to rule whether a player should be banned from their site or not. They are necessarily using a broad brush, and whoever shows up they ban.
Fair enough.
The only question is whether they should tone it down, and not publicly humiliate known individuals. |
|
 |
|
|
< Earlier Kibitzing · PAGE 17 OF 22 ·
Later Kibitzing> |
Kibitzing
