Author: Ricardo Gibert
Date: 09:19:54 08/15/03
Go up one level in this thread
On August 15, 2003 at 04:18:54, Uri Blass wrote: >On August 14, 2003 at 22:40:04, Ricardo Gibert wrote: > >>If doubling NPS (all else being equal) produces a roughly 40 to 75 elo strength >>increase, then doubling knowledge produces how much elo strength increase? >> >>Or do increases in knowledge follow a completely different type of relationship >>to playing stength? >> >>What impact on NPS does a doubling of knowledge have? I know it slows it down, >>but what is the relationship (a rule of thumb) between the 2? It is said that >>adding knowldege in hardware does not exact a speed penalty, but can this really >>be true? Or is the speed penalty simply less apparent? >> >>How much attention should a programmer give to lowering the EBF of his program >>versus adding more knowledge? > >I believe that lowering the EBF is more important. I agree with this 100%. > >> >>I'm curious about the different opinions programmers have on this topic. My >>assumption is the answers to these questions can have a big impact on computer >>chess program design. > >I believe that big evaluation is not very important for the low levels(weaker or >equal to Junior8) and it is only important to have some important cheap >things(Junior8 does not use most of its time in evaluation). > >Big evaluation may be important at higher level than Junior8 but I am still not >at that level and I do not know if there is somebody at that level. I would expect enlarging the eval would produce diminishing returns and not payoff nearly as much as people expect. Improving pruning and extentions is more important IMO. > > >I have not only a poor evaluation but a bad order of moves and poor extension >and pruning rules. >When I say bad or poor I do not compare it with other programs but with what I >expect to have in the future. > >Inspite of all these problems movei is a dangerous opponent for every program >and even the best programs of WBEC cannot get 100% against it. >Deep Sjeng1.5 could only get 2.5-1.5 in WBEC(including a loss when movei started >1.e4 c6 2.d4 d6) and Aristarch4.21 could get only 3-1(2 draws). > >I guess that the reason for the fact that movei is a dangerous opponent for >almost every program is that all programs suffer from poor pruning, poor >extensions,poor order of moves and poor evaluation. > >Uri
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.