This is a photo showing that people visiting the Yoko Ono “Imagine” exhibit in Montreal are happily following her words to “Play It With Trust”. The exhibit is moving, and it’s great to see the originals of some iconic writings, drawings, and actions!
IM Ken Regan’s own efforts to restore trust in chess are still going apace. In fact, one of his co-workers was in Nashville earlier this week presenting the first academic paper from this work at a data-mining conference: http://www.ieee-ssci.org/index.php?q=node/55 (1/5-way down page)…
3:12PM “Skill Rating by Bayesian Inference [#9060]”
Giuseppe Di Fatta, Guy Haworth and Kenneth Regan
The University of Reading, United Kingdom;
University at Buffalo, The State University of New York, United States
Thank you Ken for sending it in.
How can they tell the difference?
Thanks, Susan, for also mentioning my work along with the photo. Two other papers are currently in review, and a lot of material intended to update my “Fidelity” site is pending that and other factors—some of which are sensitive about more-recent cases. I did post a recent update there with full treatment on the Aeroflot 2009 accusation.
The papers develop statistical models that are capable of assessing skill based on the intrinsic quality of one’s moves, rather than by results of games as with Elo ratings. Then based on skill and the nature of the games, we can predict the frequency with which a non-cheating player would expect to match Rybka or any other chess engine’s moves, at any time or depth. The average expectancy for Super-GMs is 57%, but for Black’s 10 non-theory moves in the Mamedyarov-Kurnosov game, my model says any super-GM would expect to match 8-or-9 of them (it actually says “8.35”).
My real point here is completely apart from any question of guilt or “just 10 moves” (or just-3-cherry-picked games, in Mamedyarov’s update letter) being (in)significant. My point is, folks: some games will have more engine-matching just by their nature. Before saying anything rash you need to assess this nature—and as a minimum for any letter like Mamedyarov’s that mentions doing engine tests to be entertained, you need to provide log files of your tests, just as with any science experiment you did in school.
Yo Regan of all the places on the earth for Ultramodern Theory to come up do you think Buffalo would have been on the map by some random African America? heheheehe I wonder what garry thinkin right now!! haha