Quoting Dr. Regan: ""What I'm saying, as justifying my not needing to take the time to individually into tournaments to see which were broadcast and which were

Quoting Dr. Regan: ""What I'm saying, as justifying my not needing to take the time to individually into tournaments to see which were broadcast and which were not, is that if there is any bias in my data, then it's towards broadcast games (i.e more of it is analyzed due to availability) and yet I show something entirely normal." This is a simple logic error. The guy might be reputable but haven't really thought this through. Quick explanation of the mistake if it's not immediately obvious: it's possible to cheat in a way that his statistical method won't catch but the difference in performance between broadcasted and not broadcasted tournaments will highlight it. His statement assumes that his statistical model will catch every cheater which is obviously nonsense. His methods might be very good in a sense that if they show cheating we can be sure cheating in fact occurred but it can't be good enough to find every cheater, if anything because of limitation of input: just the chess moves. For example he wouldn't catch a stronger GM communicating moves to Hans (it's still very human and normal, just stronger) while comparing performance in broadcasted and not broadcasted tournaments would (if someone communicated the moves to Hans in broadcasted ones). It's very hard to take the guy seriously after that, especially when he is very arrogant about "there is no evidence" statement while it would be more prudent to say his limited statistical model didn't find any.