Computer Chess Club Archives


Search

Terms

Messages

Subject: Re: pv score oscillation

Author: Robert Hyatt

Date: 17:12:57 10/20/97

Go up one level in this thread


On October 20, 1997 at 20:12:26, Robert Hyatt wrote:

>On October 20, 1997 at 13:58:37, Chris Whittington wrote:
>
>>
>>On October 20, 1997 at 13:28:03, Willie Wood wrote:
>>
>>>
>>>On October 20, 1997 at 08:52:02, Chris Whittington wrote:
>>>
>>>>
>>>>Of course this is all gobbledegook for the fast/crafty/fritz paradigm
>>>>since they (a) don't exchange evaluate and (b) have anything remotely
>>>>like a pruning function operating over the higher plies. Gobbeldegook
>>>>for the 'programmer-programmers', perfect sense for the 'chess
>>>>player-programmers' :)
>>>>
>>>>Chris Whittington
>>>>
>>>>
>>>
>>>
>>>Bob has said he uses a SEE.  In what sense do you mean that crafty
>>>doesn't exchange evaluate?
>>
>>I don't believe he uses the swap-off function, otherwise known as SOMA
>>or exchange evaluation. This means evaluating WITHOUT the quiesence
>>search which carries out the capturing possibilities and evaluates
>>material at the end of the capture sequence.
>>
>>A swap-off, SOMA or exchange evaluator evaluates the capture sequence as
>>part of the evaluation function; and is thus prepared to terminate the
>>search without then doing a quiecence search.
>>
>>A quiecence search is what you do when you have a fast evaluator. They
>>all do it.
>>
>>Swap off evaluations are generally done (f at all) by the slower
>>knowledge based programs.
>>
>>Some programs mix and match between the two. They do swap-off
>>evaluations and terminate sometimes, or capture search sometimes,
>>depending on circumstances. Again the circumstances require knowledge
>>measurements that the fast programs can't/don't do.
>>
>>Ed has reported his capture search is around 10-15% of total nodes.
>>CSTal's is also around this figure, maybe lower.
>>
>>I don't know the precise capture search rates for the fast programs, but
>>I think their rates are very much higher than these.
>>
>>
>>Chris Whittington
>>
>>
>>
>>>
>>>WW
>
>I think I am around 30%.  The big issue between Ed and me when we had
>this
>discussion is that we were measuring different things.  I measured *any*
>node
>with depth>0 as a non-capture node, and any node with depth <= 0 as a
>capture
>node.  Ed separates the "leaf" nodes out and in that discussion counted
>them
>as non-capture nodes.  When I do that, generally 30% of the total nodes
>are
>at depth < 0...  which isn't far off from the above...



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.