29 Nov 08
1 edit
Originally posted by KeplerI`m not accusing you in manipulating of game scores. You could simply use something different from methods used by game mods without malice intention.
All the games from the tournament are available in pgn here:
http://www.chessbase.com/news/2008/games/wccc08.pgn
I could give you my copy of this but then you could not know whether I had manipulated the game scores or not. Better that you look at the original data and make your own conclusions. I selected 25 games from the 45 available by removing all th ...[text shortened]... up rate would have been much lower!) and then choosing 25 games at random from those remaining.
P.S. Also one possible reason for low matchups - in engine tournaments engines are connected with much much stronger processor that any home PC/laptop ever had.
29 Nov 08
2 edits
Originally posted by KorchOn the contrary it could be you who is leaking data from the disbanded game mod team as where is Squealbelch getting these figures from an "ex-game mod". Added to your violations of section 6 and 3c, one wonders how you got to be a game mod in the first place if you are the source of the leak. Was there a slot for the mafia in the team?
You should understand that disclosing ALL fineses of used methods will help cheats to avoid detection.....
- Joined
- 26 May 02
- Moves
- 72546
29 Nov 08
Originally posted by KeplerOk, so you chose 25 games at random out of a possible 36 to analyse. I don't understand why you did that, as it was a very small sample size to begin with.
All the games from the tournament are available in pgn here:
http://www.chessbase.com/news/2008/games/wccc08.pgn
I could give you my copy of this but then you could not know whether I had manipulated the game scores or not. Better that you look at the original data and make your own conclusions. I selected 25 games from the 45 available by removing all th ...[text shortened]... up rate would have been much lower!) and then choosing 25 games at random from those remaining.
How many games from the 1922 tournament did you analyse?
If you didn't analyse them all, then which games were discarded and what were your reasons for doing so?
29 Nov 08
2 edits
Originally posted by JieOnly reason of your accusations is your obsession 🙂
On the contrary it is could be you who is leaking data from the disbanded game mod team as where is Squealbelch getting these figures from an "ex-game mod". Added to your violations of section 6 and 3c, one wonders how you got to be a game mod in the first place.
P.S. Should say that I don`t agree with data given by Squealbelch for 100%.
29 Nov 08
1 edit
Originally posted by David TebbI analysed 25 games from the 1922 tournament. The reason I chose 25 was simply because I suffered a power cut just at the point that analysis of the 26th game had begun. Since I did not want to repeat the analysis I used 25 games from the computer tournament. Since each game is written to a separate file the cut off during the analysis of the 26th game will not have affected the previous 25 games. If the power cut had not intervened I would have analysed all 62 games and then chosen 36 at random to match with the 36 games from the computer tournament.
Ok, so you chose 25 games at random out of a possible 36 to analyse. I don't understand why you did that, as it was a very small sample size to begin with.
How many games from the 1922 tournament did you analyse?
If you didn't analyse them all, then which games were discarded and what were your reasons for doing so?
The particular test I used to determine if there was evidence that the two samples were derived from different populations works better if the samples are approximately equal in size so I went for equality. The test is valid for small sample sizes so that should not be a factor.
29 Nov 08
Originally posted by SquelchbelchSquelchbelch says 2 ex-game mods have supplied him with information that was privy to the game mod team. If anyone neutral wants to know how ridiculous this situation is it is like a forum mod releasing information to X so that he can engage in his dispute with Y. How many guesses as to the identity of the "ex-game mods"?
Dear Russ,
.......
I have recently analysed 20 games by a top player on this site and the average results exceeded these limits by quite some way.
After submitting the evidence & posting it on a private forum, I was informed by 2 people who have acted as Games Moderators on your site that the player involved had similar matchups when investigated in 2005.
It is my understanding that you were made aware of this.
......
Regards,
Steve
29 Nov 08
Originally posted by KorchI know you are not accusing me. I just want to make sure you get the original data without any possibility of alterations I am unaware of or have forgotten about.
I`m not accusing you in manipulating of game scores. You could simply use something different from methods used by game mods without malice intention.
P.S. Also one possible reason for low matchups - in engine tournaments engines are connected with much much stronger processor that any home PC/laptop ever had.
I already thought about the hardware aspect as you will see in my edited version of my reply with the link to the games. I intend to run an engine tournament on my distinctly ordinary hardware to see if that produces a difference.
- Joined
- 26 May 02
- Moves
- 72546
29 Nov 08
Originally posted by KeplerOk, thanks. Which games did you analyse though?
I analysed 25 games from the 1922 tournament. The reason I chose 25 was simply because I suffered a power cut just at the point that analysis of the 26th game had begun. Since I did not want to repeat the analysis I used 25 games from the computer tournament. Since each game is written to a separate file the cut off during the analysis of the 26th game will n ...[text shortened]... so I went for equality. The test is valid for small sample sizes so that should not be a factor.
Some more questions:
What size Hash Tables did you use? (did you use the same Hash settings for all the games?)
How did you determine when a game had gone out of "book"?
Did you use the same method for determining "book" moves in both tournaments?
29 Nov 08
5 edits
Originally posted by JieLearn to read. He refers to investigation made in 2005. I was not game mod in 2005 so I could not investigate that guy then.
Squelchbelch says 2 ex-game mods have supplied him with information that was privy to the game mod team. If anyone neutral wants to know how ridiculous this situation is it is like a forum mod releasing information to X so that he can engage in his dispute with Y. How many guesses as to the identity of the "ex-game mods"?
P.S. And I dont need to refer on earlier investigation as I`ve been investigated that guy myself few months before and results have been posted in OTB players club private forum and send with fair play ticket.
29 Nov 08
Originally posted by David TebbI analysed the following games from Vienna 1922:
Ok, thanks. Which games did you analyse though?
Some more questions:
What size Hash Tables did you use? (did you use the same Hash settings for all the games?)
How did you determine when a game had gone out of "book"?
Did you use the same method for determining "book" moves in both tournaments?
Maroczy v Kmoch, Alekhine v Reti, Koenig v Gruenfeld, Rubinstein v Takacs, Samisch v Maroiczy, Tarrasch v Vukovic, Reti v Wolf, Takacs v Tartakower, Alekhine v Samisch, Tartakower v Gruenfeld, Wolf v Kmoch, Koenig v Tarrasch, Tarrasch v Rubinstein, Vukovic v Koenig, Takacs v Alekhine, Samisch v Wolf, Gruenfeld v Maroczy
Maroczy v Bogoljubow, Spielman v Samisch, Wolf v Takacs, Rubinstein v Vukovic, Tartakower v Tarrasch, Reti v Kmoch, Gruenfeld v Wolf, Vukovic v Tartakower
I analysed the following games from the 16th World Computer Chess Championship:
Rybka v The Baron, Shredder v Junior, Cluster Toga v HIARCS, Jonny Beijing v Deep Sjeng, Cluster Toga v Rybka, Jonny Beijing v Junior, Deep Sjeng v HIARCS, Falcon v Rybka, Jonny Beijing v Shredder, Junior v Deep Sjeng, Shredder v Falcon, The Baron v HIARCS, Cluster Toga v Jonny Beijing, Jonny Beijing v The Baron, HIARCS v Junior, Deep Sjeng v Rybka, Falcon v Cluster Toga, Rybka v HIARCS, The Baron v Junior, Shredder v Deep Sjeng, Jonny Beijing v Falcon, Falcon v The Baron, HIARCS v Shredder, Junior v Rybka, Deep Sjeng v Cluster Toga.
I used 256 Mb hash table for all the games. The machine used was a dual 1.3 Ghz G4 Mac with 1 Gb RAM running Glaurung 2.1 within the Sigma Chess interface.
I also carried out a separate analysis of each game using the Sigma engine set at 5 seconds per move. The reason for this is that the Sigma engine leaves a marker when book is left. The Sigma book can be modified to reflect current opening knowledge.
Both tournaments were treated in exactly the same way.
30 Nov 08
2 edits
Originally posted by KeplerTo say that because a tournament composed of top Grandmasters 86 years ago had a match up rate (first moves only) similar to what a computer chess tournament had recently that "match up rates are no indicator of engine use" and "If anyone has been banned on the basis of match up rates alone I consider that there is at least a 50% chance they were wrongly banned. is bizarre to say the least.
You may recall I posted in another thread now long gone stating that I was applying statistical analysis to two samples of games. One sample was taken from a tournament held in Vienna in 1922 and featured the likes of Reti, Gruenfeld and Rubinstein. The other sample was taken from the 16th World Computer Chess Championship which was recently won by Rybka.
...[text shortened]... it has taken so long to ban some alleged cheats, match up rates are no indicator of engine use!
Players on RHP aren't anywhere near as good, and will never be anywhere near as good, as top GMs of ANY era. This is true even taking into account the advantages of correspondence play, which most players here don't take full advantage of (they play too many games too fast). It's not very shocking that Rubinstein usually made very strong moves that agree with an engine's recommendations; it is kinda shocking when 1200 RHPers all of sudden start doing so.
I don't know what a better indicator of someone using an engine is then that his moves are what an engine suggests. Of course, there are other factors to consider but once a match up rate exceeds that which the best correspondence players of all time were able to achieve in far less games using far more time, then "beyond reasonable doubt" is satisfied. This of course assumes a significant sample size.
30 Nov 08
1 edit
Originally posted by no1marauderI don´t think that you can make an assumption that all players on RHP are rubbish as a basis for banning them when they start winning. Although I take your point - what could be done is have some known honest players play some games on a site which allow engines and compare those games with their games here to test the system. I think that that would be a fairer test.
To say that because a tournament composed of top Grandmasters 86 years ago had a match up rate (first moves only) similar to what a computer chess tournament had recently that "match up rates are [b]no indicator of engine use" and "If anyone has been banned on the basis of match up rates alone I consider that there is at least a 50% chance they we yond reasonable doubt" is satisfied. This of course assumes a significant sample size.[/b]
On a point of pedantry a sample is not in itself significant. The result is what is significant. You can have a huge sample and still not get a significant result, and have a small sample which is nevertheless significant. It depends on the size of the effect you are trying to measure. In the jargon they talk about the power of a sample, which depends on sample size and (expected) effect size.
Edit: Had to add a [ /b ] to stop Marauders post causing mine to come out bold.
30 Nov 08
Originally posted by DeepThoughtNo one said or assumed all players on RHP are "rubbish", did they? But none are quite as good as Rubinstein or Reti, are they? Statistical studies have been done that give a reasonable outer limit for what match ups can be achieved by good human players on RHP without cheating. More tests are unnecessary.
I don´t think that you can make an assumption that all players on RHP are rubbish as a basis for banning them when they start winning. Although I take your point - what could be done is have some known honest players play some games on a site which allow engines and compare those games with their games here to test the system. I think that that would b ...[text shortened]... effect size.
Edit: Had to add a [ /b ] to stop Marauders post causing mine to come out bold.
Thanks for the nitpicking. You and anybody else who has followed this issue (though you actually only started following it), know what I meant.
30 Nov 08
Originally posted by no1marauderThere's the rub & I don't see how anyone can argue against matchup rates as the principle evidence in deciding to ban or not to ban a user accused of engine use.
...I don't know what a better indicator of someone using an engine is then that his moves are what an engine suggests...
I can see the disputes now:
"Oh yes his engine move matchup out of book & over time was only 30% for top 3 choices but to me it looks like he was playing like an engine"
and
"While he does matchup with Fritz's top 3 matches out of book in 30 games 98% of the time, to me his play looks just like a tactically gifted human player"
Lol
30 Nov 08
The word "You" isn't directed at any individual.
I don't understand the level paranoia that some have when it comes to on-line games. If you tend not to trust anyone you can't see then it's probable that you think that everyone who defeats you somehow cheated or got lucky. The best thing to do is play OTB and frisk your opponent there for the hidden pocket computer when you lose.
If you've faded to the point where you think there are secret programs watching what's running on your machine, it's time to get a check-up at the local mental health clinic. Any decent firewall or security program would pick it up in a heart beat, block it and notify you.
I've thought that a few have played much over their rating when they trounced me, but after thinking it over settled knowing they probably just had a real good game. I'm actually winding down playing the game for awhile but at this point I consider playing on-line the same as playing a computer opponent on Chessmaster, and have no expectations other than learning and getting better. If people put half of the effort proving that someone cheated on them into reading and learning some more, maybe they'd get a little better. If you don't trust someone don't play them. If you don't trust anyone, don't play here or any other internet site, You'll feel alot better.
You know who you are.
😀