Author: Keith Evans
Date: 18:19:20 07/02/02
Go up one level in this thread
On July 02, 2002 at 20:42:16, Russell Reagan wrote: >Which computer chess tournaments have hardware limitations and which are >"anything goes"? > >To me, a tournament between computers where "anything goes" is meaningless. If >it is a computer vs. human, then that is something entirely different I think. >What does taking a super computer to a tournament and winning against (possibly >superior) engines running on slower hardware prove? To me it doesn't prove >anything other than you had the money to win a tournament. To me that doesn't >imply that any engine was better than another if it's anything goes. IBM could >build another super computer and run an alpha-beta search with piece-square >table evaluation and win the "world championship", and it hasn't proved that it >was the best engine. > >Is anyone else turned off my a competition between computers where it's open >hardware? I don't think it proves a thing as far as which engine is better. I >guess it depends which aspect you are interestd in. If you're interested in >hardware, then you probably like the open hardware competitions. If you're >interested in AI in computer chess, then you're probably more likely to enjoy an >equal hardware competition. To me a competition that can be bought doesn't mean >anything. It might as well just be a bidding process to see who is the next >"champion". > >Any thoughts? > >Russell Well I'm a hardware engineer and the open competitions are quite interesting to me. It brings variety. If open hardware just meant overclocking then it would not be interesting to me. But by forcing everybody to use exactly the same system you would miss out on some novel ideas. Let's say that Brutus performs quite well and is eventually commercialized. Would you be interested in purchasing it at the right price? It seems to me that it really belongs in the open competition category, and that without that opportunity to compete it might never be developed. Would Vincent have been motivated to support massive parallel processing if he couldn't use it in competition? What if he makes some interesting chess discoveries as a result? (And are you really absolutely convinced that Vincent is going to be the winner of this competition?) One thing that you might do is to construct a graph of platform cost versus performance for past competitions and see how correlated they are. I'm sure that you'll see some very expensive platforms that didn't fare all that well. Of course I also think that it's nice to have competitions with standardized platforms. And yes the element of human competition (regardless of platform) would also make it more interesting to me. Regards, Keith
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.