Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: Tuning? Rubbish!

Author: Michael Cummings

Date: 03:56:15 12/16/99

Go up one level in this thread


On December 16, 1999 at 03:39:48, blass uri wrote:

>On December 16, 1999 at 02:56:56, Michael Cummings wrote:
>
>>On December 15, 1999 at 19:54:08, Christophe Theron wrote:
>>
>>>On December 15, 1999 at 18:14:27, Michael Cummings wrote:
>>>
>>>>On December 15, 1999 at 13:43:19, Christophe Theron wrote:
>>>>
>>>>>On December 15, 1999 at 08:45:39, Ulrich Tuerke wrote:
>>>>>
>>>>>>> ...
>>>>>>>That is what I include in tuning an engine. Plus when more games are played
>>>>>>>against tiger programmers can tune their engines not just towards tiger but a
>>>>>>>bit towards all programs, an engine upgrade.
>>>>>>
>>>>>>"Tuning" against all opponents ?
>>>>>>I would call this "improving" instead of "tuning".
>>>>>
>>>>>
>>>>>RIGHT!
>>>>>
>>>>>
>>>>>    Christophe
>>>>
>>>>Which is what I said, an engine upgrade. Tuning an engine, is when you upgrade
>>>>it. Did I not write it that way ? You tune an engine you are upgrading the
>>>>quality of it, right ?
>>>
>>>
>>>Weren't you talking about tuning against a specific opponent or against a
>>>specific subset of opponents?
>>>
>>>
>>>
>>>    Christophe
>>
>>
>>I was talking about tuning the engine, which includes the learning feature for
>>programs that have it. And tuning the engine so when it has played aginst
>>programs a certain amount of times and the programmers see an area where it can
>>improve it tunes the engine towards that area. Which is basically an engine
>>upgrade.
>>
>>I will be interested to see now that other programs will be able to play tiger
>>how long it stays were it is. People were very happy to say CM6K would be dead
>>because it has no learning feature and that it would be about 5th or 6th best on
>>the 200MMX machines. Well it is still No 1 on the 200MMX machines, this is after
>>programs with learning functions have had a chance to learn, and still cannot
>>get ahead of it.
>
>I guess that it is number 1 on the 200MMX because there is no tiger on p200MMX.
>if you assume the same difference relative to Fritz532 then tiger is number 1
>on p200MMX
>
>Uri

Does that mean I can assume that CM6K is number 1 on the 450 machines ?
Well tiger might be number 2 behind CM6K on the 200MMX machines, and maybe CM6K
on a 450 machine will play way better than anyone expected on the 450 hardware.

You can assume that way Uri, but I will assume my way. Alls fair after all,
right ?



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.