Author: Robert Hyatt
Date: 19:58:42 01/23/00
Go up one level in this thread
On January 23, 2000 at 19:25:55, Tom Kerrigan wrote: >On January 23, 2000 at 19:20:14, Jeremiah Penery wrote: > >>>If you can give me an exact description of DB's eval function, I will pay you >>>$100 (possibly more), and I will release a program using that function in less >>>than a month. >>We don't have an exact description, but we do know a lot of what it does. Why >>aren't _any_ of the programs now doing certain things that DB _was_ doing? > >Like what? Please, just give one example. I mentioned one from Hsu's book: "potentially open files". I've never heard of anyone doing it. I have never seen any example of anyone appearing to do it. In one of the games in the last match, Hsu (in his book) mentions that a rook move (maybe R to the a file somewhere) was made because the eval termed the a file a "potentially open file" even though the search couldn't see it opening anywhere. This was something that I think Benjamin suggested. I will try to find the exact game/move he mentioned tomorrow when I get to the office... There are other things dealing with king safety and pieces attacking around the king, through other pieces and pawns, etc.. that are extremely expensive to compute. _at the tips_. > >>>I suspect that CS Tal's evaluation function does more work than DB's. I actually >>>have evidence. >>So, what is the evidence? > >Hsu's estimate of 40k instructions per node. > >>> And so far, nobody's been able to prove me wrong. Also, there's >>>no reason why you can't use the DB eval function with a search function that's >>>more appropriate for a PC. >>Then it wouldn't quite be DB, now would it? > >No, but my original suggestion was just to implement the DB evaluation function, >which is supposed to be spectacular. Over the course of this argument, I have >suggested in passing that it would be possible to implement the entire DB >algorithm on a PC. I never thought this was a particularly good idea, but I >definitely think it's possible. > >-Tom That is where we disagreed. I absolutely said it was not a good idea. And could see _no_ reason Hsu would invest that much work. I know _I_ don't write code just for the heck of it... And I absolutely hate to rewrite already existing code for the heck of it... which is what we would be asking him to do: "Hsu, please take your hardware design, reproduce it in C, and put it into a program so we can see how it does. We know it will be way slower. And that your old search won't work well as it was tuned for a much faster program. And that the eval weights might not be tuned to the new shallower depth... But do it anyway." Berliner reported in a HiTech paper that when he tried testing at very shallow depths it broke his program because his eval assumed a certain basic search depth to find some simple tactics... and when he ran the "hitech vs lotech" tests to try to predict rating per ply increases, he saw this. I don't think it is possible to just re-do DB in a PC disguise. I think Hsu would start over and end up with something pretty similar to what everybody else has. Evolution has not brought us all to the same 'neighborhood' accidentally...
This page took 0.01 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.