Author: Brian Richardson
Date: 10:39:04 10/10/00
Go up one level in this thread
On October 10, 2000 at 11:12:17, Mike Adams wrote: >Hi pulsar currently uses R2 (depth-3) nullmove untill maxdepth 7 search then it >kicks in at R3. I noticed that R3 seems to give more depth. Pulsar also uses >futility pruning in search. if material value plus possible capture plus >futility value for that depth <alpha move is not searched. I'm not sure this >futility pruning is so sound so what i was thinking was get rid of pruning in >search and be more aggressive with null move. What are peoples opinions >particularly programmers? I'll probably test this somewhat but its hard to get >games against high rated opponents for pulsar so i'm hoping for some direction >before the testing. These are all very good questions. I have experimented with various combinations for Tinker and the options tend to work well (or poorly) with each other in interesting ways. This in combination effect makes testing key. For nullmove, I suggest the ICCA write ups on adaptive nullmove. Tinker uses R=3 at depth > 6 but also checks for more than 2 pieces left to try and avoid endgame zugzwang problems.
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.