Author: José de Jesús García Ruvalcaba
Date: 15:38:20 02/11/00
Go up one level in this thread
On February 11, 2000 at 17:58:25, Vincent Vega wrote: >On February 11, 2000 at 11:23:13, blass uri wrote: > >>I think there is an assumption that slow searchers and fast searchers have the >>same branching factor and also that you get the same elo increase from 5 plies >>to 6 plies and the question if you are a fast searcher or a slow searcher is not >>relevant. >> > >You're correct of course, but I never meant that the increase in strength would >be the same for different programs with the increased ply depth. When comparing >the increase in strength across the programs, we should exponentially increase >the evaluation time. > >>I think that the last assumption is wrong because slow searchers simply can see >>more tactics at the same depth because a big evaluation can tell you things >>about tactics(for example that there is a fork). >> >>Uri > >Yes, the falloff (if there is one) would occur at different ply depths for >different programs. But, I believe the results from one program showing the >falloff will mean that all programs using min-max will experience the falloff >somewhere - not necessarily at the same depth. I think that these results would >mean that both slow searchers and programs with big branching factor would have >an advantage in the future. I do not know if slow searchers will have an advantage in the future (due to the faster hardware) or not. But I believe that programs with a big branching factor will have a big disadvantage in the future. I think that the most succesful programs with faster hardware will have a very good search policy. José.
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.