Author: Christophe Theron
Date: 10:20:23 02/24/03
Go up one level in this thread
On February 24, 2003 at 12:10:51, Chessfun wrote:
>On February 24, 2003 at 11:58:57, Mogens Larsen wrote:
>
>>On February 24, 2003 at 11:46:21, Albert Silver wrote:
>>
>>>I have to agree with Rolf. She should only publish her results until there is a
>>>statistically reliable amount of games, otherwise people might be misled into
>>>believing Fritz 8 is actually 1 Elo point stronger than Deep Fritz 7.
>>>
>>> Albert
>>
>>The problem isn't only the number of games, where the accuracy is covered by the
>>SD information. But also the game collection presents some problems, which are
>>partly mentioned, so the list is purely "for fun". That is why I categorized it
>>as "not" being reliable in my previous comment and it doesn't presume to be. Not
>>that hard to understand, really.
>>
>>Regards,
>>Mogens
>
>I will in future rename the Fun List the Unreliable Fun List.
>This should go some way to satisfy.
>
>Sarah.
Just rename it "Reliable Chessfun list for those who understand statistics and
Unreliable Chessfun list for those who don't".
I find your work valuable and correct from a scientifical point of view, but
it's probably because I understand statistics a little bit.
Looks like some people here don't, so they probably need a good book on the
subject.
No matter how many times the subject has been discussed, it seems that every
time a new list is published we need to explain all over again what a margin of
error and a confidence interval is.
It's not the "Computer Chess Club" anymore. It's the "Statistic For Dummies
Club" now.
Welcome to SFDC.
Christophe
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.