Oct

5

I recently sent this letter to Chessbase after seeing the latest Bulgarian attack:

Dear Chessbase,

It was with interest that I read Mr Danailov's so-called 'statistical analysis' of the number of Kramnik's moves coinciding with the recommendations of Fritz 9. I am not an expert in statistics by any means, but in this case I might be able to offer some advice.

In order to show that the number of moves coinciding with Fritz 9's recommendations in this match is of a certain 'significance', Mr Danailov should test Kramnik's moves outside the match and then assess whether the 'match subset' differs by several standard deviations. In this way he might establish 'confidence' limits.

 There has been some interesting work in this area by Steve Levitt, author of 'Freakonomics', who the Bulgarian team might like to consult (no doubt at huge expense) if they are sincere in their claim. And to demonstrate their integrity and even handedness they might also commission a similar study of Topalov's moves, both during the match and after his ascendency to the number one spot.

I would expect that both studies would be entirely consistent with randomness, but who knows?

Sincerely, Nigel Davies (International Grandmaster of FIDE)


Comments

Name

Email

Website

Speak your mind

Archives

Resources & Links

Search