Author: Christophe Theron
Date: 15:09:24 10/14/03
Go up one level in this thread
On October 14, 2003 at 11:29:50, Gerd Isenberg wrote:
>On October 14, 2003 at 03:46:40, Christophe Theron wrote:
>
>>On October 13, 2003 at 14:56:57, Gerd Isenberg wrote:
>>
>>>On October 13, 2003 at 14:19:14, Christophe Theron wrote:
>>>
>>>>On October 13, 2003 at 13:09:03, Charles Roberson wrote:
>>>>
>>>>>
>>>>> You make the statement that Diep is a positional engine and you chose it based
>>>>>on that. So, why did you run G/5 matches? At G/5 tactics and search depth
>>>>>is crucial.
>>>>
>>>>
>>>>
>>>>I would like to bring to your attention that tactics and search depth are
>>>>crucial at any time controls in chess.
>>>>
>>>>Showing dimishing returns from increased search depth is so difficult that in
>>>>practice there is little difference between blitz and long time controls.
>>>>
>>>>If engine A gets a beating at blitz, expect it to get the same beating if you
>>>>repeat the match with long time controls.
>>>>
>>>>
>>>>
>>>> Christophe
>>>
>>>Christophe,
>>>
>>>couldn't it be, that engines have some odd/even sympathy/antipathy?
>>>If the characteristic line of this property is to be out of phase between two
>>>programs, i can imagine that one is a better blitzer but the other the better
>>>medium time player.
>>
>>
>>I did not say that it is impossible to build very unbalanced chess programs.
>>
>
>Yes, but the question is the degree of (un)balance. Is the
>performance-searchdepth(time) characteristic graph a straight line or more or
>less discontinuous around a dimmish line or curve?
I tend to think it is a monotonic, asymptotic function.
In other words there is no maximum in this curve.
It is possible to build a pathologic chess program that would have a strange
curve, however. But I don't see any existing interesting chess program that has
this problem.
Christophe
>>>And what about fast against slow with "more" or "better" knowledge. At blitz
>>>time control the the linear speedup (fast/slow) may be more important. But due
>>>to superior branching factor at sime time the "better" knowledge pays off and
>>>the match tilts.
>>
>>
>>* Superior branching factor can be achieved even with a high NPS.
>>
>>* fast/slow vs dumb/knowledged is an outdated conception. Fast does not mean
>>dumb and slow does not mean knowledged. It is an overly simplistic view of the
>>problem that should be definitely forgotten.
>>
>>
>>
>> Christophe
>
>
>I agree that the extremes became indistinct, but i believe
>the issue is still valid to some degree, even in your league.
>
>Cheers,
>Gerd
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.