Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Ignore post above, Great idea! would be interested in a faster version

Author: Rémi Coulom

Date: 23:49:09 07/17/04

Go up one level in this thread


On July 17, 2004 at 17:35:50, Russell Reagan wrote:

>On July 17, 2004 at 15:53:04, Robin Smith wrote:
>
>>would be interested in a faster version
>
>Steve Maughan used to have a program available that would work only with the
>result of a match and give similar info. You enter the loser's score and the how
>sure you want to be (in percent) about the result, and it tells you the required
>number of wins. You can enter the loser's score and play with the percent until
>the number of wins is correct, and get the confidence percentage that way.
>Unfortunately it appears the content on his website is missing at the moment. It
>used to be here.
>
>http://www.stevemaughan.com/downloads/whoisbetter.exe
>
>I have the program, and I uploaded it here if you are interested.
>
>http://home.comcast.net/~r.reagan/WhoIsBetter.exe
>
>Here is a screenshot of the program, to get an idea of how it works.
>
>http://home.comcast.net/~r.reagan/whoisbetter.jpg

This program does not solve the same problem. I also have a similar program in
my statistical significance tools, that is fast and accurate. Also note that
Steve's program has a little bug. Results are off by one unit. For instance, for
your screen shot, 13-5 gives a confidence of 96.8%, but 12-5 gives 95.2%. So
only 12 wins are required to ensure 95% confidence with 5 losses.

Rémi



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.