Author: Robert Hyatt
Date: 06:44:56 02/17/00
Go up one level in this thread
On February 17, 2000 at 03:17:33, Tony Werten wrote: >On February 16, 2000 at 13:20:03, Robert Hyatt wrote: > >>>Just don't expect to see it play those 10 programs. >>> >>>Tony >> >>Exactly what would this be based on? Deep Thought dominated computer chess >>events for 10 years. DB is _much_ stronger. What would lead you to think that >>the 'best 10 programs in the world could beat it' now, when they couldn't do it >>for 10 years running? DT lost 2 games that I know of against microcomputers >>over a 10 year period. The micros lost plenty of games against DT. DT searched >>about 2M nodes per second, while Deep Blue was 100X faster, not to mention >>better. > >DB jr didn't search 200M . According to the artikel it searched 20M/s. > >My idea is that most computersprograms have improved more over the last 10 >years. >The hardware has definately improved more than 10X. But even besides that, also >the programs itself have been improved. > >Another ( subjective ) argument: Have you seen DB jr play in a computerevent >lately ? > >Tony DB Jr was a 1997 program. And yes, it played in several human events thru 1997 to prepare for the real Kasparov match. I watched it play several games at a couple of SuperComputing conferences. DB Jr was just a 'piece' of DB. But 20M nodes per second simply pales the speed the rest of us are getting on the fastest hardware of today. I am doing over 1M on my xeon. I am still 20x slower. And I don't do anywhere near the evaluation things they do because I can't afford them. I see nothing at all that suggests that the best programs of today could beat them. DT won the first fredkin prize. DB Jr is two generations beyond deep thought. And far stronger.
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.