Author: Enrique Irazoqui
Date: 10:38:27 02/09/00
Go up one level in this thread
On February 09, 2000 at 13:33:33, blass uri wrote: >On February 09, 2000 at 13:16:59, Enrique Irazoqui wrote: > >>On February 09, 2000 at 13:02:30, blass uri wrote: >> >>>On February 09, 2000 at 12:54:24, Enrique Irazoqui wrote: >>> >>>>On February 09, 2000 at 12:50:57, blass uri wrote: >>>> >>>>>On February 09, 2000 at 12:25:08, Enrique Irazoqui wrote: >>>>> >>>>>>On February 09, 2000 at 12:00:36, blass uri wrote: >>>>>> >>>>>>>2rqr1k1/p3bp1p/6p1/n1pnP3/3pB1b1/PP3N2/2QN1PPP/B1R1R2K b - - 0 1 >>>>>>> >>>>>>>This position is from game 19 between Fritz6a and rebel Century. >>>>>>>Rebel Century played 22...Rb8 after 81 seconds and lost the game. >>>>>>> >>>>>>>My original Rebel Century >>>>>> >>>>>>That's it. Did Ed send you the patched version that played in my tournament? >>>>>>It's supposed to be identical to yours, except for the removal of some code that >>>>>>improved the speed by 8%. It had to do with a discussion here between Ed and >>>>>>Bob. Later on, Ed found out that this new code harmed Century, removed it and >>>>>>sent me the patch. >>>>>> >>>>>>>(not 1.2) found after less than a minute on >>>>>>>pentiumIII450(slower than enrique's hardware) the move >>>>>>>22...Nf4 and does not change its mind. >>>>>> >>>>>> >>>>>> >>>>>>>probably an autoplayer problem. >>>>>> >>>>>>Nope, I just checked. Century picks Rb8 at ply 7.02 after 15 seconds, +0.38 eval >>>>>>and keeps it until ply 9.01, when it switches to Nf4 after 238 seconds, +0.43. >>>>>>Ed can confirm. >>>>> >>>>>I played a game against the original Rebel century that I got when I bought it >>>>>and it found Nf4 in less than a minute. >>>>> >>>>>I used the infinite time control and told it to move when it used enough time. >>>>>It could produce all the moves before Nf4 but swithced to Nf4 after less than a >>>>>minute. >>>>> >>>>>Maybe the difference is because I used the game and not the position without the >>>>>game. >>>> >>>>I used the game too, of course. >>>> >>>>>I do not know. >>>> >>>>If you check it with the patch Ed sent me, you will get the same results. >>>> >>>>Enrique >>> >>>It seems that this patch is not productive for Rebel Century >> >>Well... what can I say. If Ed sent it for the tournament it must be because it >>is at least not worse than the original Century. > >I guess that he thought that it is not worse but the question is if he was right >> >>> because the >>>commercial Rebel century without the patch can find Nf4 faster(only 58 seconds >>>on my pentiumII450) >> >>One move, come on.. It doesn't prove that a program is better than another. > >My Rebel gives exactly the same score for Rb8 at depth 7(0.38) >and exactly the same score for Nc4 at depth 9(0.43) but is wise enough to >changes its mind at depth 8 for Nc4 (evaluation 0.41) Uri, do you really think that a difference of 3/100 of a pawn in the evaluation is enough to tell which version of Century is better? Ed's criteria deserve more credit than that, don't you think? >I used 60Mbytes hash tables and strongest setting. So did I. Enrique >Uri
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.