Author: Robert Hyatt
Date: 08:40:15 11/26/02
Go up one level in this thread
On November 26, 2002 at 07:21:12, David Rasmussen wrote:
>On November 26, 2002 at 05:22:42, Dann Corbit wrote:
>
>>
>>Are you absolutely sure that the nodes are being counted correctly? This is not
>>exactly a wild-open sort of position with scads of deep combinations in waiting.
>> It's hard to imagine how the qnodes can go berserk like that.
>
>I'm pretty sure I'm counting them correctly. I count only the nodes that
>Q-search _generates_, not the node I'm in on entering Q-search. Normally, my
>Q-percentage isn't that high. Before I got SEE, it was regularly in the 60%'s.
>Now, it is usually about 20-40%. But in this particular position (and the ones
>"near" it), and probably many others, the Q-search explodes.
>
>I'm counting like this:
>
>Score Search(pos,...)
>{
> ++nodes;
> ...
> if (depth < ONE_PLY)
> {
> --nodes;
> return Quiescence(pos,...);
> }
> ...
>}
>
>and
>
>Score Quiescence(pos,...)
>{
> ++nodes;
What is the purpose of that? Each recursive call to q-search incs nodes
which doesn't make sense to me.
> ...
> while (anymore interesting moves)
> {
> ...
> MakeMove(pos,move);
> score = -Quiescence(pos,...);
> UnMakeMove(pos,move);
>
> ++qNodes;
> ...
> }
>}
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.