Author: Jeremiah Penery
Date: 12:10:52 01/21/00
Go up one level in this thread
On January 21, 2000 at 13:51:06, Alvaro Polo wrote: >On January 21, 2000 at 13:38:21, blass uri wrote: > >>It is not clear to me that the evaluation function of deeper blue was a real >>advantage relative to crafty's evaluation(deeper blue failed to expect Qe3 in >>the game that kasparov resigned so there are cases when crafty's evaluation is >>better and if deeper blue could see one or two more plies thanks to using >>simpler evaluation than it is another advantage). > >If I have understood correctly DB evaluation was done in hardware and was >essentially "free", meaning that if replaced by a simpler evaluation the number >of cycles wouldn't be reduced and it wouldn't search any faster because of this. Exactly. If I had such a capability, I'd put everything I could possibly think of into the eval. Attacks, square control, x-rays, pins, mobility, tempo (if possible), pawn structures (passed pawns, pawn majorities, pawns around king, etc.), king safety (including a bunch more attack-around-king information, pawn shelter, open files, half-open files, Rooks/Queens attacking, batteries, etc.), overloaded pieces, weak pieces, etc. I know that much of this stuff is in DB, and I'm guessing about other parts. A couple of these things probably aren't in it. All this AND a monster tactical ability...And some people think the micros could possibly compete?!
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.