Page Index Toggle Pages: 1 ... 5 6 [7] 8 9 ... 12
Topic Tools
Very Hot Topic (More than 25 Replies) Yelena Dembo on Chess.com (Read 99770 times)
ChevyBanginStyle
Full Member
***
Offline


2 \infty & *CRUNCH*

Posts: 238
Joined: 01/03/10
Re: Yelena Dembo on Chess.com
Reply #87 - 09/22/10 at 04:09:03
Post Tools
drogo, if you're representing chess.com, you're certainly not doing them any favors. Frankly, you are simply ignorant about correspondence played at a high level. And it's not like grandmasters have not tried their hand at international correspondence. (It's not a walk in the park.) I don't consider myself an expert on correspondence chess, and I imagine a lot of people here could educate you on the matter much better than I could; that is, if you were willing to learn. 

You are simply wrong about all the important novelties coming from the top in OTB chess. Many "novelties" are "discovered" by a top OTB player and only then reach widespread attention. These correspondence masters don't play correspondence chess for fame and fortune. They play it for love of the game... or maybe they play for bananas, since they are monkeys.
  
Back to top
 
IP Logged
 
drogo
YaBB Newbies
*
Offline


I Love ChessPublishing!

Posts: 40
Joined: 09/21/10
Re: Yelena Dembo on Chess.com
Reply #86 - 09/22/10 at 02:03:44
Post Tools
There is nothing wrong with ICCF: the rules are clear, computers are allowed and you can play if you accept this. Rules are also clear on chess.com: no engines are allowed and, if you don't accept this, you can play on ICCF. There are two different games, you choose the one you like to play. Or you can play both, no problem! But it is unfair to use engine assistance on a site which explicitly forbids that and then come with justifications like "it is normal in CC".

Nobody challenged my analysis, nobody provided an example of a human player who achieved similar engine correlation. It's much easier to sit in an armchair and say "the method is flawed". 

Ads are an IQ test on the internet. Some people pass it by googling "ad block". Some people never pass it and instead complain Smiley

As far as I know the top rated Gm on chess.com is Julio Becerra and he is 300 points below Yelena. He played like 150 games.

Markovich, shall I make a drawing to help you understand that I AM NOT ACCUSING ANYONE OF CHEATING??? And I have no idea what are the criteria of chess.com? All I'm trying to do is to understand what are the limits between human players and engine users.
  
Back to top
 
IP Logged
 
Markovich
God Member
*****
Offline



Posts: 6099
Location: Columbus, Ohio
Joined: 09/17/04
Re: Yelena Dembo on Chess.com
Reply #85 - 09/22/10 at 01:29:09
Post Tools
Smyslov_Fan wrote on 09/21/10 at 19:06:56:
In drogo's defense, if we can't distinguish between human and computer play, then the top echelons of CC play will indeed be inhabited only by those who combine their skills with chess engines. Several players who have challenged chess.com's assertions about WGM Dembo here agree with this premise.  Most of them believe that there can be no such test. 


This is nothing in drogo's defense; it is a non sequitur.  It bears nothing upon the question of the adequacy of the methods in question, which remain secret, and which have been challenged here without any adequate rejoinder. 

By this logic, if reading Tarot cards were the only available method of guessing who were cheating, the paramount purpose of purifying correspondence chess would justify accusing people based on Tarot readings.

Start another thread if you want to wring your hands over the purity of correspondence chess.  If you do, I will remark there that leading forum for it, ICCF, tolerates computer assistance, thousands of players participate there, and no one seems the worse for it or anywhere near as exercised about it as you are. 

drogo wrote on 09/22/10 at 00:40:21:
ChevyBanginStyle wrote on 09/21/10 at 21:54:21:


Serious correspondence chess is for researchers. Researchers use tools to make their discoveries. Efficiency is a major aspect of productive research in a competitive setting. Computers are practically a necessity for competitive research (i.e. serious correspondence chess) nowadays.


You call them researchers, I call them monkeys.


My, my!   It turns out that we do have a chip on our shoulder!  Send a letter to ICCF, why don't you?   Mount your stallion and start your crusade!  But if you do, I suggest you arm yourself with a better weapon than the obscure numerology so far discussed in this thread.

One thing I am quite sure of, though, and that is that chess.com with its anonymous players, crappy interface and required viewing of flash ads is unlikely to replace www.iccf-webserver.com anytime soon.

P.S.  On chess.com Yelena Dembo outrates all other titled players by 200-300 elo?  Truly?

drogo wrote on 09/21/10 at 14:08:57:


I am not aware of a public set of data. It's likely that the big names in online chess like ICC have their own statistics, but they won't make it public. Mainly because they don't want cheaters to adapt their cheating methods.


Well then, the logic of the accusers becomes rather circular, doesn't it?  Rather like that used at Salem, where those accused as witches weren't permitted to know the evidence against them for fear they would cast spells on their accusers.  Or like Joe McCarthy, who never revealed his list of names of "known Communists" in the State Department, lest this alert them to his investigation.

Say, I wonder if there are any "known Communists" on chess.com?
  

The Great Oz has spoken!
Back to top
 
IP Logged
 
drogo
YaBB Newbies
*
Offline


I Love ChessPublishing!

Posts: 40
Joined: 09/21/10
Re: Yelena Dembo on Chess.com
Reply #84 - 09/22/10 at 00:40:21
Post Tools
ChevyBanginStyle wrote on 09/21/10 at 21:54:21:


Serious correspondence chess is for researchers. Researchers use tools to make their discoveries. Efficiency is a major aspect of productive research in a competitive setting. Computers are practically a necessity for competitive research (i.e. serious correspondence chess) nowadays.


You call them researchers, I call them monkeys. The great opening novelties are discovered by the OTB players. All the chess books are written by OTB players. When Anand and Topalov prepared for their match, spending so much time and resources on opening preparation, they chose their seconds among the OTB players. Damned, they both ignored those CC "researchers"!
  
Back to top
 
IP Logged
 
Smyslov_Fan
God Member
Correspondence fan
*****
Offline


Progress depends on the
unreasonable man. ~GBS

Posts: 6902
Joined: 06/15/05
Re: Yelena Dembo on Chess.com
Reply #83 - 09/21/10 at 23:16:40
Post Tools
Chevy, your position is certainly reasonable. But consider this, it is not the position that Yelena Dembo has taken. Dembo has denied using a computer to assist her.
  
Back to top
 
IP Logged
 
ChevyBanginStyle
Full Member
***
Offline


2 \infty & *CRUNCH*

Posts: 238
Joined: 01/03/10
Re: Yelena Dembo on Chess.com
Reply #82 - 09/21/10 at 21:54:21
Post Tools
Smyslov_Fan wrote on 09/21/10 at 19:06:56:
In drogo's defense, if we can't distinguish between human and computer play, then the top echelons of CC play will indeed be inhabited only by those who combine their skills with chess engines. Several players who have challenged chess.com's assertions about WGM Dembo here agree with this premise.  Most of them believe that there can be no such test.


I thought computer use in correspondence chess at the highest levels was generally accepted. If you try to impose artificial rules in situations were computers are heavily used for opening research, you'll always have these problems. Dembo wrote a repertoire book on the Gruenfeld which is a theoretical opening with many forcing lines.

Is chess.com a website for correspondence chess at the highest levels? I thought it was clear it was not. It's heavily commercial and I think they are making a dumb business decision that will drive away titled players.

Quote:
I hold out hope that we can develop some sort of Turing test for correspondence chess. I still find chess.com's methodology interesting. Unfortunately, they apparently weren't sufficiently rigorous in their testing. It's impossible to tell until they release their full method of testing.

I hope chess.com heeds the growing calls to be fully transparent in its methodology. I also hope that titled players will help the websites to develop methods to catch cheats without falsely labelling strong players.


Serious correspondence chess is for researchers. Researchers use tools to make their discoveries. Efficiency is a major aspect of productive research in a competitive setting. Computers are practically a necessity for competitive research (i.e. serious correspondence chess) nowadays.

Honestly, I admit this is not my cup of tea (at least not yet). My main goal is to become a strong OTB player, and I don't want to be too dependent on engines in analysis. I realize that many approach chess in a different way and I accept that. It does not hurt my ego that they can use their research abilities to play chess at a very high level in correspondence. Actually I respect that.

Personally, I think attempting to write a "Turing Test" for correspondence play is futile as long as chess continues to advance. Furthermore, I don't think it's the people who are currently playing correspondence chess at the highest levels who have these concerns. Either you quit or you learn to adapt. There will always be people who would rather play without computers. I don't think it's too hard to find such people.
  
Back to top
 
IP Logged
 
Smyslov_Fan
God Member
Correspondence fan
*****
Offline


Progress depends on the
unreasonable man. ~GBS

Posts: 6902
Joined: 06/15/05
Re: Yelena Dembo on Chess.com
Reply #81 - 09/21/10 at 19:06:56
Post Tools
In drogo's defense, if we can't distinguish between human and computer play, then the top echelons of CC play will indeed be inhabited only by those who combine their skills with chess engines. Several players who have challenged chess.com's assertions about WGM Dembo here agree with this premise.  Most of them believe that there can be no such test. 

I hold out hope that we can develop some sort of Turing test for correspondence chess. I still find chess.com's methodology interesting. Unfortunately, they apparently weren't sufficiently rigorous in their testing. It's impossible to tell until they release their full method of testing.

I hope chess.com heeds the growing calls to be fully transparent in its methodology. I also hope that titled players will help the websites to develop methods to catch cheats without falsely labelling strong players.
  
Back to top
 
IP Logged
 
ChevyBanginStyle
Full Member
***
Offline


2 \infty & *CRUNCH*

Posts: 238
Joined: 01/03/10
Re: Yelena Dembo on Chess.com
Reply #80 - 09/21/10 at 17:35:52
Post Tools
drogo wrote on 09/21/10 at 14:54:47:
Yes, the problem is blurry, yet websites are forced to take decisions in these conditions. Because otherwise only cyborgs will play online chess.


Really? The leaps to the conclusions you make are really quite astounding.
  
Back to top
 
IP Logged
 
drogo
YaBB Newbies
*
Offline


I Love ChessPublishing!

Posts: 40
Joined: 09/21/10
Re: Yelena Dembo on Chess.com
Reply #79 - 09/21/10 at 17:25:22
Post Tools
Some of you are too quick with the conclusions. If matching an engine move by move at depth 17 is a "clear indication" of human play, it might be the case that nobody is cheating in online chess. That is possible, although strange.

For all the known human players (including the world champions) and for all the known CC players from before 1980, an engine at depth 17 points out many inaccuracies and even blunders. So, if the pattern can be confirmed over more games, I'd say that Yelena plays differently from any human we know.

I'd also be interested to know what is your analysis at depth 23.
  
Back to top
 
IP Logged
 
Schaakhamster
God Member
*****
Offline


I Love ChessPublishing!

Posts: 650
Joined: 05/13/08
Re: corr on Chess.com
Reply #78 - 09/21/10 at 16:48:12
Post Tools
TalJechin wrote on 09/21/10 at 15:20:10:


Btw, after seeing their list http://www.chess.com/forum/view/community/chesscom---list-of-caught-cheaters, which is mostly aliases; another thing occurred to me: what if a cheater chooses someone's real name as an alias? Obviously, publishing that name would not hurt the cheater but only an innocent third party. 

So, the whole "publishing for shame" should at least require that players sign up under their real name and proves their identity - but then we come to issue of different persons sharing the same name, which is not all that uncommon...


Truly an obnoxious system. 

The screenshot of Dembo's account I saw looked like http://www.chess.com/members/view/MrMagnificient 

"publishing for shame" does not really affect anonymous users hiding behind for instance the nickname MrMagnificient. So I can only imagine that the list was created to deter cheaters (ha!) or more likely to show the other users that they were tackling the main problem casual corr servers have: cheating by engine use.

  
Back to top
 
IP Logged
 
Schaakhamster
God Member
*****
Offline


I Love ChessPublishing!

Posts: 650
Joined: 05/13/08
Re: Yelena Dembo on Chess.com
Reply #77 - 09/21/10 at 16:30:02
Post Tools
drogo wrote on 09/21/10 at 14:22:36:
Keano wrote on 09/21/10 at 14:16:52:
depends on the game surely? All sounds very iffy to me this whole thing. Surely Dembo is still entitled to sue them for defamation/loss of earnings as the damage is already done now? If it was me I'd be going after them all guns blazing.


Right now that account is closed and there is no mention of cheating accusations. It's just that a site chose to close the account of one user, for whatever reason they might had. End of story.


I saw a screenshot of the website where cheating was mentioned (it was on her wiki-page, but has been removed). So chess.com changed that, probably because they smelled a rat and realized that lawyers would have a field day with the original statement. And several blogs have reported the "cheating" accusation.  So not really the end of the story. 

I don't challenge the right of chess.com using statical methods to detect probable engine-users and to suspend players based on those methods. But you just can't label some-one a cheater based on those methods alone, certainly if they are not willing to share the methods used or the results they obtained. 



  
Back to top
 
IP Logged
 
TalJechin
God Member
*****
Offline


There is no secret ingredient.

Posts: 2892
Location: Malmö
Joined: 08/12/04
Gender: Male
Re: corr on Chess.com
Reply #76 - 09/21/10 at 15:20:10
Post Tools
drogo wrote on 09/21/10 at 14:22:36:
Keano wrote on 09/21/10 at 14:16:52:
depends on the game surely? All sounds very iffy to me this whole thing. Surely Dembo is still entitled to sue them for defamation/loss of earnings as the damage is already done now? If it was me I'd be going after them all guns blazing.


Right now that account is closed and there is no mention of cheating accusations. It's just that a site chose to close the account of one user, for whatever reason they might had. End of story.


By publishing her name as a "known cheater" chess.com has started a malicious rumour that 'thanks' to the internet won't go away. Anyone searching for her name +chess, will find lots of blog entries, forum discussions etc - stuff that will never fully disappear again. And google's associative functions may add her name among synonyms to 'cheater', which then can reappear in unpredictable ways...

So this is similar to a witch-trial, where drowning in the well was the only way to prove one's innocence...


Quote:
Engine used: Firebird 1.0, multiline mode (3 lines).
Depth: 17
database used: chesslive.de


17 ply - really?!? In an iccf game I wouldn't trust any engine suggestion below 23 ply, at least. If she really only played according to a 17ply standard, I'd take that as a clear indication that she wasn't cheating.


Btw, after seeing their list http://www.chess.com/forum/view/community/chesscom---list-of-caught-cheaters, which is mostly aliases; another thing occurred to me: what if a cheater chooses someone's real name as an alias? Obviously, publishing that name would not hurt the cheater but only an innocent third party. 

So, the whole "publishing for shame" should at least require that players sign up under their real name and proves their identity - but then we come to issue of different persons sharing the same name, which is not all that uncommon...
  
Back to top
 
IP Logged
 
drogo
YaBB Newbies
*
Offline


I Love ChessPublishing!

Posts: 40
Joined: 09/21/10
Re: Yelena Dembo on Chess.com
Reply #75 - 09/21/10 at 14:54:47
Post Tools
Markovich wrote on 09/21/10 at 14:40:54:
We have people here, including you, saying that she cheated.  Furthermore it remains their policy to post a list of "known" cheaters.  So not end of story by a long shot.

From what I can tell, Regan does not put the problem in a statistical form.  So you can search his writings forever and not find answers to the critical questions.


I never said that Yelena Dembo cheated. I don't know!! All I said is that, in the game I analyzed, she matched the top choice of Rybka consistently. I am only interested in developing a clear methodology to separate human players from cheaters.

Yes, the problem is blurry, yet websites are forced to take decisions in these conditions. Because otherwise only cyborgs will play online chess.
  
Back to top
 
IP Logged
 
Markovich
God Member
*****
Offline



Posts: 6099
Location: Columbus, Ohio
Joined: 09/17/04
Re: Yelena Dembo on Chess.com
Reply #74 - 09/21/10 at 14:40:54
Post Tools
Not end of story, because they have left the impression that she cheated.  We have people here, including you, saying that she cheated.  Furthermore it remains their policy to post a list of "known" cheaters.  So not end of story by a long shot.

From what I can tell, Regan does not put the problem in a statistical form.  So you can search his writings forever and not find answers to the critical questions.

Further my results for kingboy-Dembo, obtained with DeepShredder 12, differ considerably from yours.  I have Dembo playing suboptimally at moves 9 (Bxb2 was better), 13 (Nc3 was better), 15 (Qe7 was better), 18 (Ne6 was better). 20 (Bc2 was better), 28 (Bxf3 was better), and 34 (f5, Qb1 and Ng5 were better).  In the latter case there is a pretty big difference between 34...f5! and Dembo's move.

So if we have Dembo playing any obviously suboptimal moves at all in this supposedly very convincing game, what is your account of her conduct?  That she was a ruthless cheater who sometimes forgot to cheat?  Isn't it much more likely that this IM simply played strong moves and sometimes faltered?

Furthermore, this is a highly tactical game.  There are many cases where the move played by Dembo is the only one that a strong human being would consider (moves 10, 11, 12, 14, 17, 24, 26, 27, 29, 30, 31, 32, 33 and 35).  So this "evidence" is really quite insubstantial.

If cheating were a criminal offense and 20 games like this one were your evidence, you think you could go into court and get a conviction?

Has anybody written an objective procedure such as a Python script for automatically constructing these 1-2-3-order matchups with some engine after such-and-such amount of time spent in contemplation?  For selecting which moves are subject to inquiry and which are not?  Unless they have, the whole thing smacks of ad-hockery and non-replicability.  So does the whole problem of what engine to use.

If you let your engine think for say 2 minutes per admitted move and you have a test sample of say, 5,000 games, do you have any idea of how long it would take to construct the threshold numbers chess.com uses (which remain of unknown statistical character, I again point out, but which presumably have some relation to a test sample)?  Has anyone actually done that, and based on how many games?

Really it's just ridiculous.  It looks like a load of frummery; an exercise in numerical mysticism..
  

The Great Oz has spoken!
Back to top
 
IP Logged
 
drogo
YaBB Newbies
*
Offline


I Love ChessPublishing!

Posts: 40
Joined: 09/21/10
Re: Yelena Dembo on Chess.com
Reply #73 - 09/21/10 at 14:22:36
Post Tools
Keano wrote on 09/21/10 at 14:16:52:
depends on the game surely? All sounds very iffy to me this whole thing. Surely Dembo is still entitled to sue them for defamation/loss of earnings as the damage is already done now? If it was me I'd be going after them all guns blazing.


Right now that account is closed and there is no mention of cheating accusations. It's just that a site chose to close the account of one user, for whatever reason they might had. End of story.
  
Back to top
 
IP Logged
 
Page Index Toggle Pages: 1 ... 5 6 [7] 8 9 ... 12
Topic Tools
Bookmarks: del.icio.us Digg Facebook Google Google+ Linked in reddit StumbleUpon Twitter Yahoo