Tired of hearing of GM Vladimir Kramnik controversy, I eventually decided to get to the bottom of the facts.
I can feel for those getting in the cross hair of chess anti-cheating systems mythology.
I do suggest realizing it is quite challenging, if not possible for science today to design a statistical system capable of detecting high level player computer aid.
How would one expect to assemble a reference Titled Tuesday tournament dataset to ferret out outliers?
The only strict statistical method possible is to analyze a known 100% cheat free dataset vs. one which not.
Short of such reference dataset what remains is approximation and hypothesis.
Dr. Kenneth Regan model being published can be gamed and occasional GM level computer aid, if done with parsimony, falls out of detectable pattern.
GM Vladimir Kramnik, however caustic and even reckless the behavior, has been the trigger for academics and chess.com to admit the limitations and shortcomings of those models.
I suggest weighting in playing on GM Khismatullin’s account and unprecise statistical analysis with Vladimir acting as the whistleblower which lead to improvement and renewed focus on the issue.
I continue to remember GM Vladimir Kramnik as the 2002 tie 4-4 vs. Deep Fritz 7.5, a great hope for humanity, only to be outdone in 2006 by Deep Fritz 10.
This is the first world champion vs. computer match I know to satisfy the requirement the entire machine being visible and facing the human during the match.
I hope the reader will agree variance is powerful enough an ally to deflect and refute any suspicion.
Do not defend or refute unless on the chess board!
Tired of hearing of GM Vladimir Kramnik controversy, I eventually decided to get to the bottom of the facts.
I can feel for those getting in the cross hair of chess anti-cheating systems mythology.
I do suggest realizing it is quite challenging, if not possible for science today to design a statistical system capable of detecting high level player computer aid.
How would one expect to assemble a reference Titled Tuesday tournament dataset to ferret out outliers?
The only strict statistical method possible is to analyze a known 100% cheat free dataset vs. one which not.
Short of such reference dataset what remains is approximation and hypothesis.
Dr. Kenneth Regan model being published can be gamed and occasional GM level computer aid, if done with parsimony, falls out of detectable pattern.
GM Vladimir Kramnik, however caustic and even reckless the behavior, has been the trigger for academics and chess.com to admit the limitations and shortcomings of those models.
I suggest weighting in playing on GM Khismatullin’s account and unprecise statistical analysis with Vladimir acting as the whistleblower which lead to improvement and renewed focus on the issue.
I continue to remember GM Vladimir Kramnik as the 2002 tie 4-4 vs. Deep Fritz 7.5, a great hope for humanity, only to be outdone in 2006 by Deep Fritz 10.
This is the first world champion vs. computer match I know to satisfy the requirement the entire machine being visible and facing the human during the match.
I hope the reader will agree variance is powerful enough an ally to deflect and refute any suspicion.
Do not defend or refute unless on the chess board!