Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: EVIDENCE That Junior REALLY DID Perform Very Badly In Bilbao

Author: Michael Yee

Date: 07:33:30 10/13/04

Go up one level in this thread


On October 13, 2004 at 07:38:13, Graham Laight wrote:

>Junior's performance in Bilbao was much worse than even I had thought - and now
>I have the weapon to prove it - the simulator in
>http://www.talkchess.com/forums/1/message.html?391362
>
>Starting with the assumption that all 3 machines had equal win/draw/loss
>probablities, I tried different combinations of probabilities to produce the
>kind of results that Hydra and Fritz got (3.5/4). I found that setting the win
>probablity to 70%, the draw probability to 20%, and the lose probability to 10%
>gave a 27.4% probability of producing 3.5 points - which is reasonable.
>
>However - the probability for Junior's score (1.5/4) comes out at only 2.02%!!!
>
>To those who thought that my remarks about Junior's poor performance were
>statistically illiterate, I would ask you to run the simulator, and try to come
>up with win/draw/lose probabilities that give a good chance of the results we
>saw in Bilbao happening if each computer were equally as good.
>
>I await your probability numbers...
>
>-g

Like Joachim, I don't think most people disagree with your initial observation
that Junior "underperformed" in the tournament. However, what set people off is
when you asked "why" in your initial post--as if something inherent in junior
caused it to score poorly in this particular tournament. Most people here seem
to attribute junior's relatively lower score in this particular tournament to
random effects.

Furthermore, considering the probability of an event after it occurs leads to
funny conclusions sometimes. For example, I could walk over to a parked car,
look at its license plate number and think to myself, "Wow, the probability of
me seeing XYZ123 is practically zero, but here I am looking at it!" Junior might
have 1 "bad" (or underperforming) tournament out of 20, i.e., with low
probability. But the rare event *will* (or could) happen at some point.

Michael



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.