Author: Ryan B.
Date: 15:19:03 10/20/05
Go up one level in this thread
On October 20, 2005 at 17:44:10, Bertil Eklund wrote: >On October 20, 2005 at 13:26:32, Thomas Logan wrote: > >>Anyone know how Shredders position learning works >> >>questions, wuestions, questions >> >>Is it updated every game ? >> >>Does it continue getting larger and larger ? >> >>Is it ever reset ? >> >>Is it opponent sensitive ? >> >>Tom > >I can´t answer your question but a few programs like Shredder and Nimzo for >example plays different in the same opening (after the book) and seldom repeats >the same moves even if it was a win (hash influence?). Occasitonaly I have seen >the same opening 3-4 times in the same match but Shredder or Nimzo often are the >first to pick another line. > >Bertil I am sure position learning is stored hash values. I would assume it is something like if score drops over margin after a made move to the next store the new value of the move to be preloaded into hash next game. What I don't know is how the depth or hash aging is handled. Maybe someone knows of a paper in position learning?
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.