Author: José de Jesús García Ruvalcaba
Date: 11:26:54 07/14/00
Go up one level in this thread
On July 14, 2000 at 13:06:59, Baldomero Garcia, Jr. wrote: >1. First of all, congratulations to the Deep Junior team for doing >so well in this tournament. By human standards, Deep Junior *performed* >well enough for a GM norm. However, I don't think FIDE's rules allow >computers to score GM norms or get GM titles, let alone be rated by FIDE. Actually, FIDE will not rate any game from a tournament with computer entries, even if played between two humans. Silly decision, very FIDE-like. >2. I guessed from the beginning of the tournament that Deep Junior would >score 50% in this tournament. Well, the tournament hasn't ended, but DJ >has been around 50% most of the tournament. I think the main reason is >because the GMs have been afraid to even try to win. Kramnik is the only >one who attempted to win and actually did. Of course, if the GMs played >with a bit more aggression they might also lose, but I still think they've >been very timid. I think that the big hardware really pays off even against humans. I have read that there are diminishing returns and that faster hardware is more helpful against other computers, but definitely it always helps. >3. In the last game we saw, between Anand and DJ, I think Anand kind of >tried to copy Kramnik, but instead of going for the Stonewall, he went for >the Colle, and Anand didn't get anything out of the opening. I do not understand why these grandmasters do not use their home-prepared opening novelties against the machine. Perhaps they overestimate the computer's book because it is big, but it is one of the program's weakest points. >4. Question: when is the next tournament in which a computer plays against >humans? I do not know of any scheduled tournament yet, but if you want more games I can play a match against Amy (: Or, if you want the human to win, I can play against TSCP. > I know another program (I don't know much about it) PConner or >something like that, is doing quite well in another tournament. > P.ConNerS stands for "Parallel Conspiracy Numbers Search". It uses a conspiracy numbers search (CNS), a search algorithm radically different from the popular principal variation search (PVS). It runs on a massively parallel computer. Do a web search to find more about it. >Baldomero Garcia 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.