On Apr 25, 4:46 am, David Richerby
wrote:
raylopez99 wrote:
http://www.chessbase.com/newsdetail.asp?newsid=3455

This was the article I was thinking of, per my earlier post, not the
Jeff Sonas article.

That article is, frankly, junk: I'm surprised it was ever accepted for
an academic conference.

I am in the process of reading this article now, and just
noticed a laughably absurd claim by the authors: that
the truncated Crafty used would naturally rank all
superior programs in reverse order. LOL! This is the
most ignorant comment I have seen since before I began
ignoring many recent postings by the Evans ratpack.

Of course, it is not the strength, but rather the
*similarity in style*
which would actually determine how truncated Crafty
ranks *all* other programs. It is theoretically possible
for Crafty to rank Rybka near the top, penalizing it
(unfairly) only for the few moves which it correctly sees
but where Crafty would blunder horribly. All this would
require is that Rybka *usually* agree with Crafty, but
when they disagree, for Rybka to always be right. The
gap in ratings could easily be 400 points, if the key
differences of opinion were instant game-losers.

I am beginning to get the impression that people who
play around with statistics in an attempt to demonstrate
something, are loony, as well as utterly incompetent in
applying statistics rationally.

-- help bot