Author: Vincent Diepeveen
Date: 19:04:34 01/09/00
Go up one level in this thread
On January 09, 2000 at 12:55:20, Christophe Theron wrote: >On January 09, 2000 at 10:57:52, Vincent Diepeveen wrote: > >>On January 09, 2000 at 10:52:02, Enrique Irazoqui wrote: >> >>>On January 09, 2000 at 10:30:32, Vincent Diepeveen wrote: >>> >>>>On January 09, 2000 at 10:26:20, Enrique Irazoqui wrote: >>>> >>>>>On January 09, 2000 at 10:17:58, Vincent Diepeveen wrote: >>>>> >>>>>>On January 09, 2000 at 09:52:57, Enrique Irazoqui wrote: >>>>>> >>>>>>>On January 09, 2000 at 09:47:56, Thorsten Czub wrote: >>>>>>> >>>>>>>>On January 09, 2000 at 09:43:21, Enrique Irazoqui wrote: >>>>>>>> >>>>>>>>>No. Without opening book and after 1.e4 Tiger picks: >>>>>>>>> >>>>>>>>>Ply 1 Nf6 >>>>>>>>>Ply 3 d5 >>>>>>>>>Ply 3 e6 >>>>>>>>>Ply 5 e5 >>>>>>>>>Ply 8 d5 >>>>>>>>>Ply 12 e6 >>>>>>>>>Ply 13 e5 >>>>>>>>> >>>>>>>>>Enrique >>>>>>>> >>>>>>>>good that you confirm enrique. so it is on my machine too. >>>>>>>>good. >>>>>>> >>>>>>>Any time. :) >>>>>>> >>>>>>>If it happened, it happened. This need to confirm is ridiculous. Imagine if you >>>>>>>have to replay every game... The 40 games you play are already quite something, >>>>>>>no? >>>>>>> >>>>>>>Enrique >>>>>> >>>>>>But diep plays without book now and with a 9 ply search or something >>>>>>you won't play the opening good! >>>>> >>>>>Well... At least Diep played 5 or 6 moves in book and Tiger none... >>>>> >>>>>In any case, Diep should play with the book it is supposed to use, but this >>>>>checking of Tiger moves is unnecessary. >>>> >>>>what's wrong with that? >>>>Tiger plays everywhere in the world 1..d5 but what was sent to thorsten >>>>plays opening a lot better suddenly. it's tuned for playing without book >>>>there! >>> >>>Really! :) >>> >>>Look, Vincent, I already answered to this in another post. With the parameters >>>in the ct.ini file that Chrsitophe recommends for Rebel-Tiger 12.0e, Tiger picks >>>e5. After 1.e4 and with the old parameters, it picks d5 from ply 9 on, 13'', >>>without switching to e5. That explains it, doesn't it? By the way, these >>>parameters are recommended for all games played by Tiger, not only for this >>>match. >>> >>>What's wrong with double checking moves one by one? That it is physically >>>impossible to recheck all moves in 40 games, so you better believe in the good >>>faith of the whole thing. Besides, the purpose of this match was to prove that >>>Tiger's evals "suck" and that it wins because of the book. Now, if it wins >>>without a book as it did, how can we describe the evaluation functions of the >>>program that loses to Tiger? >> >>no program understands opening. if i put a 9 ply search versus a 13 ply >>search in a position where they both understand shit from, >>then i know who wins. > > >Vincent, do you realize that you are simply implying here that your program is >going to be unable to compete against Tiger? >I can believe you when you say that Diep does not understand shit from the >opening, but please don't assume Tiger doesn't know how to play the opening >without book. christophe force your silly program to 1..d5 be a man. see how big blunders it makes then as it doesn't have a weak pawn at e5 to put its pieces too. 1.e4,d5 2.ed5,qxd5 3.nc3 and then it plays like Qd6?? forever. after that d4 Nf6 Nf3 a6? Not i'm the man covering up things. you're the man covering up things! So many positions it plays bad! > Christophe >>If i play on a quad xeon with diep in that opening nothing goes wrong >>though. 12..14 ply is what i call for years the tactical barrier. >> >>Note that diep plays 6.O-O at 12 ply and keeps that move and would >>have gotten 12 ply at a quad xeon. >> >>If you call me crazy that i removed pruning from diep (otherwise >>it would also get 12 ply here and play O-O), just do so. >>I'm used to parallellism nowadays... >> >>I would love to compare evals indeed in normal positions. Not at move 6 >>where you need to get to the tactical barrier. >> >>>Enrique >>>>>Enrique
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.