Author: Christophe Theron
Date: 09:29:31 06/15/02
Go up one level in this thread
On June 15, 2002 at 11:56:23, Uri Blass wrote:
>On June 15, 2002 at 11:40:30, Christophe Theron wrote:
>
>>On June 15, 2002 at 07:11:48, Thorsten Czub wrote:
>>
>>>On June 13, 2002 at 23:58:46, Christophe Theron wrote:
>>>
>>>>On June 13, 2002 at 09:13:43, Robert Henry Durrett wrote:
>>>>
>>>
>>>>
>>>>
>>>>The problem is that the problem described above does not exist.
>>>>
>>>>
>>>>
>>>> Christophe
>>>
>>>ok -
>>>
>>>a human beeing has a different velocity than a guy on a bike.
>>>and a different velocity than a guy in a car.
>>>
>>>we have a gepard
>>>we have a guy running,
>>>a guy on a bike
>>>and a guy in a sports car.
>>>
>>>the race begins.
>>>
>>>they all have different behaviour.
>>>
>>>after 3 meters
>>>after 15 meters
>>>after 50 meters
>>>after 100 meters
>>>after 1000 meters
>>>
>>>the same counts for chess programs.
>>>
>>>put fritz, hiarcs and shredder
>>>on a pc.
>>>
>>>let them play 5' games
>>>10 minute games
>>>40/120.
>>>
>>>and you get the same different behaviour.
>>>
>>>
>>>if all chess programs would behave linear concerning strength,
>>>if all creatures would behave linear,
>>>it would easier to understand.
>>>
>>>but the creatures and the programs do not speed up linear.
>>
>>
>>
>>What you are referring to is the "branching factor".
>>
>>From the results at fast and long time controls it is possible to see if some
>>programs have better branching factors than others.
>>
>>As far as I can tell, among the top commercial chess programs (including Hiarcs
>>8) you will not find notable differences.
>>
>>
>>
>> Christophe
>
>There are some problems here.
>
>I guess that it is easy for you to have better branching factor by more
>aggresive pruning but your program is going to be weaker.
>
>You can also ask what is the branching factor of Junior.
>
>Unclear because if it jumps from depth 12 to depth 15 then it has a bigger
>branching factor than the case that it jumps from depth 12 to depth 14.
>
>Uri
Yes that's right.
I guess that the only way to measure the combination of "search efficiency" and
"branching factor" is by comparing results at very different time controls.
In this I factor out the differences due to evaluation because I don't think
evaluation makes a significant difference depending on the time controls (it
contributes approximately to the same amount of "elo" disregarding the time
controls).
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.