Author: Dann Corbit
Date: 16:08:04 01/30/02
Go up one level in this thread
On January 30, 2002 at 19:03:58, Andrew Dados wrote: >On January 30, 2002 at 19:00:28, Dann Corbit wrote: > >>On January 30, 2002 at 18:55:41, Andrew Dados wrote: >>[snip] >>>Do you really believe they searched 4 nodes *only* at first ply? >> >>Probably not. >> >>>18 ply with bf of 4 is x*4^17, where x is somewhere from 40 to 1000 depending on >>>how many nodes qsearch needs. >>> >>>for 32 nodes ridiculous minimum (say - no qsearch) it will be: >>>32x4^17/200M = 46 minutes to complete 18 ply. >> >>From what starting position? Are we picking the most difficult position on the >>board? >>Here is an 18 ply search I did yesterday with Chess Tiger: >>[D]8/8/1p1r1k2/p1pP2p1/P3KnP1/1P3N2/8/3R4 w - - ce 72; acd 18; acs 600; pv Ne5 >>Rd8 Nc4 Re8+ Kf3 Ne2 Rf1 Nd4+ Kg3+ Kg7 d6 Rd8 Rf2 Rd7 Rh2 b5 axb5 Nxb5 >>Rh5;|Nf3e5 >> >>It took all of 5 minutes. >> >>It seems that we are assuming that we *must* start with the most difficult >>imaginable position. > >No, but CT has bf of 2,5 to 3. 2.5^18 = 14551915.228366851806640625 4^18 = 68719476736 4^18/2.5^18 = 4722.366482869645213696 Is it inconceivable that Deep Blue is 4722 times as fast as Chess Tiger? Everyone who is boggling at 18 ply is making unwarranted assumpions. Of course, if it really is from the most difficult possible position and we only have 3 minutes, and it is a brute-force fullwidth search with no pruning of any kind whatsoever, then I don't believe it either.
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.