Author: Miguel A. Ballicora
Date: 10:26:12 01/30/02
Go up one level in this thread
On January 30, 2002 at 11:15:03, Gian-Carlo Pascutto wrote: >On January 30, 2002 at 11:04:18, Vincent Diepeveen wrote: > >>On January 30, 2002 at 10:12:25, Gian-Carlo Pascutto wrote: >> >>>Why are you blindly assuming that their effective branching factor >>>is 40, just because that is the average number of legal chess >>>moves? You should know better. And I know you do. >> >>branching factor is not the issue here GCP and you know it. >>the total number of nodes needed when assuming a theoretical >>minimum here is more important. >> >>Please get your college book again and see that this is about >>the squareroot of the number of legal moves. > >The formula is: > >nodes = (branching factor)^floor(depth/2) + (branching factor)^ceil(depth/2) -1 > >Your formula is an approximation. > >>The average number of legal moves, not counting checking positions as >>those get extended, it is 40. >> >>so squareroot(40) is what you need for nodes. > >For an alphabeta with perfect move ordering and NO enhancements >whatsoever, yes. But that is not what Deep Blue is. I tested >a comparable version of Sjeng. >I got the number 10-20 out of experimentation, rather than out of my a**. > >>>They used PVS. Aspiration windows. Hashtables for the first TWELVE >>>plies. Even futility pruning. >> >>In 1998 and 1999 it was mentionned by direct postings from Hsu >>and others that they only did a fullwidth search not a single form >>of pruning as they disbelieved this, Bob has quoted that zillions of >>times in these years. > >Bob'll have to be the reference on this, but I always understood they >didn't use any form of pruning except for quiescent futility pruning, >which is what I tested with. > >>>The _worst_ I saw was around 20, on _average_ it was only about 10 or >>>even less. >> >>that would make it 18^20 then in your case = 9^400, >>where i used 18^squareroot(40) = 9^40, get the point? > >Your maths is wrong. So wrong, I suspect you and Bob >attended the same school :) What's a young guy on one the Boston's supermarkets waiting on the fast lane for "10 products or less" with 11 products? two choices 1) a Harvard student that can't count 2) a M.I.T. student that can't read :-) Maybe they did not attend the same school :-) Let it go guys, these Deep Blue threads are starting to look like Bobby Fischer's thread or more like Elvis Presley's threads. Regards, Miguel, who can't read or count but get the right answer for the wrong reason all the time :-) > >-- >GCP
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.