Author: Stephen A. Boak
Date: 13:09:25 01/03/00
Go up one level in this thread
Hi Graham,
Evidence is evidence, if it is shown. Then its worth may be judged by the
observer. If it is not shown, any results that may be quoted are merely
opinions, submitted second hand at best.
Are the calculations used by Selective Search to arrive at its results openly
available (including test conditions, game scores and results)? I am referring
specifically to that which backs your recollection of published findings
indicating the difference between chess software ratings at normal (tournmanent)
play, Active (+80 pts) and Blitz Chess (+200 pts).
Selective Search admits on its current website pages that comp-comp and
comp-human results may be very different. Yet it says it calculates a chess
software rating that combines the two types of results!
This hodge-podge rating for the best rated program on its list is one you have
mentioned and 'relied on' to some degree to make a point in one of your above
posts.
If I recall correctly, Selective Search does not publish the underlying data
and show its calculations for its published chess software rating list. Am I
correct?
I doubt that the +80 and +200 point conclusions you mentioned are fully
documented as well. Do you know otherwise? Are those figures also based in
some fashion on the 'hodge-podge' ratings of Selective Search?
I like mathematics. I like using it as a tool to measure the world around us.
But I don't think we should rely (for evidence) on calculations that have not
been shown in the light of day.
I am not saying your memory is inaccurate. I am not saying the Selective
Search calculations are worthless. I am not saying they were not carefully
compiled. I am saying that without divulging the underlying test conditions,
results and subsequent calculations, we have no real evidence other than a
published 'result', bare of content, that must be treated as mere opinion.
Respectfully (so you don't take this the wrong way!),
--Steve Boak
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.