Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Let's talk about fraud.

Author: Rolf Tueschen

Date: 09:09:50 05/03/04

Go up one level in this thread


On May 03, 2004 at 11:14:59, Vincent Diepeveen wrote:

>On May 03, 2004 at 03:47:41, Rolf Tueschen wrote:
>
>>On May 03, 2004 at 01:50:44, Gian-Carlo Pascutto wrote:
>>
>>>On May 02, 2004 at 21:03:06, Robert Hyatt wrote:
>>>
>>>>But it is within the noise
>>>>range for non-deterministic behavior anyway...
>>>
>>>I showed that the difference was statistically
>>>significant, so that argument holds no water.
>>
>>Statistical significance isn't simply a little calculation, you must also care
>>about the correct statistical setup; how can you believe that you could
>>contradict the noice theory for non-deterministic behavior with a little math
>>calcul? Please tell us exactly what you did under what conditions and then what
>>you've found. Step by step so that we could follow.
>
>don't quote nonsense. hyatt shows 1 or 2 positions and very seldom 3 or 5
>positions to proof something and then claims 3.1 speedup out of 4.
>
>any test done so far over a lot of positions by anyone else than hyatt using his
>own program, even at his own machine, shows way less than 3.1.
>
>2.8 in fact.
>
>Anyone denying statistical significance for measuring speedups has a hole in his
>head.

I have exactly 5 holes in my head. Now let's count the significance. I deny it.






>
>>
>>
>>>
>>>--
>>>GCP



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.