Re: Using partial plys

["P.J.Leonard" <P.J.Leonard@xxxxxxxxxx>]

David Fotland wrote:
> 
> Many Faces uses best first search without depth limitation for
> life and death searching.  (well, actually there is a depth
> limitation as I implemented it, but lines can end at any depth,
> and most lines don't reach the depth limitation.)
> 
> It's not proof number search, since PNS only uses the number of
> reasonable move at each node to guide the search. 

 Actually the information is the number decendent nodes that need to
be proved in order to prove the node in question. It also uses the
number of
nodes required to refute the node. The initial proof/refute guess can
use heuristics.
I also inceremental expand nodes which can be used to add the most
promising
nodes first. This does something to get around the problem you are
hinting at.

> My move generator
> has much better move confidence values that it uses instead.

  Is this published ?  What information is stored for each search node
and how is this information used to update ancestor nodes and direct the
search. 
Sometimes it is better to explore a less promising node if you know it
will only
require a little work to prove/refute compared promising nodes that
would require much work
to prove because there are so many oponent responses to refute.


-- 
 Cheers Paul.                                        
               __         __                  
              (  ) _ - _ (  )                 
               ~~         ~~                  
                (  0   0  )                   
----------ooO----(       )----Ooo------------ 
          '''     (     )     '''             
                   (   )                      
                    \ /                       
                   ~~O~~                      

  Tel: +44 1225 826108
  Fax: +44 1225 826305
snail: P.J.Leonard
       Applied Electromagnetic Research Centre
       Bath University,
       BATH, UK   BA2 7AY