Author: Gian-Carlo Pascutto
Date: 23:48:16 02/19/02
Go up one level in this thread
On February 19, 2002 at 23:27:00, Robert Hyatt wrote: >OK if you talk about null-move I assume _nobody_ says that it doesn't add >its own unique errors into the search. I certainly see enough of them. :) Perhaps. But you are still using it... >No... but a failed pruning rule can and will eliminate a critical move >that skews the scores badly. The more errors you make, the more the >score gets skewed. An error at ply=2, for example, is simply critical >and will result in a very bad error. It should be the nature of a good pruning rule that such a mistake will alost never happen. i.e. that it is more reliable near the root This is true for most rules in use now, either by making the decision based on more data (nullmove) or simply by not using it near the root (futility pruning) >by definition any forward pruning is bad, in terms of introducing error. >Yes, you might gain more than you lose with some rules. But trying to >make the tree selective enough to reach that 60 ply repetition is certainly >going to push the error rate into the stratosphere.... This is very much a Vincent-ian argument. It's not because you can't do it that it is simply impossible. For one, Chess Tiger seems to be searching nowadays with a branching factor that's very close to 2 and sometimes even below that. It's obviously not getting any kind of huge error rate that causes it to lose each game by playing losing blunders. Far from it, even. I see no reason not to believe this can't be pushed further down. Now think about it. Once you consistently go below 2, each simple speedup will result in more and more plies. You won't need to see the full 60 (or 36) ply in every variation here. The draw by repetition is going to have a lot of checks in it which should allow a program to find it faster. I'm assuming a 30 ply nominal search might suffice. That is not so far out as you might think. Seeing it's at 22 ply after half a day, it'll need (2^8) / 2 = 128 days as a rough estimate. >It has been around forever (version 1 and 2 and 3) version 1 was R=1, non- >recursive. Version 2 was R=2 recursive. Version 3 was R=2~3 I currently >use it. Version 4 might be pure R=3 as some are trying... But they are >all very closely related... > >Whether there will be a revolutionary new approach is a guess... I hope so, >but I doubt it... I would say it has already arrived. Those darned pro's are just not sharing it :) -- GCP
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.