Author: Uri Blass
Date: 11:29:30 06/28/03
Go up one level in this thread
On June 28, 2003 at 11:26:43, Uri Blass wrote: >On June 28, 2003 at 10:39:34, Drexel,Michael wrote: > >>On June 28, 2003 at 09:59:30, Uri Blass wrote: >> >>>On June 28, 2003 at 08:12:37, Aaron Gordon wrote: >>> >>>>4 seconds Slate. Does this mean Goliath Light 1.5 is better than Deep Blue? No. >>>>Does this mean Deep Blue is better than Goliath Light 1.5? No. Does this mean >>>>Goliath Light 1.5 is better for this particular position? Yes. >>>> >>>>:) >>> >>>Or maybe it means that GoliathLight is lucky to overevaluate the queen and it >>>helps it in this position. >> >>I dont think Goliath is lucky here, it is simply a strong tactician. >> >>> >>>The main problem of engines that cannot solve the problem fast like movei that >>>may need some hours is problem to see that white is significantly better after >>>the following moves: >> >>I dont think so. The main problem is to find 3.Bg7. Once an engine finds this >>move it gives positive scores very quickly. > >Finding 3.Bg7 is easy for movei but the score is simply smaller than the score >for 1.Qh3 so it has not this line at the root position. > >If I give it Rxe6 then it has no problem to have Bg7 in the main line in a short >time but seeing a score that is better than the score for Qh3 takes a very long >time so I am convinced that the problem of not finding 1.Rxe6 in an hour is >result of positional reasons. I see when I analyze it again that it is relative and the short time that Movei needs to see 3.Bg7 in the main line after Rxe6 is less than a minute but still more then what little goliath needs before 1.Rxe6(my programs analyzed other things and I could not check the exact short time then. I still believe that the main problem for programs that cannot solve it in some minutes is not Bg7. I may change the value of the queen to check if movei can solve it in some minutes. 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.