Author: Uri Blass
Date: 15:10:38 07/08/03
Go up one level in this thread
On July 08, 2003 at 17:33:27, Alessandro Damiani wrote: >On July 08, 2003 at 10:40:28, Uri Blass wrote: > >>On July 08, 2003 at 10:28:55, Djordje Vidanovic wrote: >> >>>On July 07, 2003 at 16:47:09, Uri Blass wrote: >>> >>>>On July 07, 2003 at 16:44:28, Djordje Vidanovic wrote: >>>> >>>>>Here's the clip from the newest Ruffian engine: >>>>> >>>>>New game >>>>>4rqn1/p3brk1/R5p1/6Bp/1p5Q/1B1R4/PPP2PPP/6K1 w - - 0 1 >>>>> >>>>>Analysis by Ruffian 2003-06-23: >>>>> >>>>>1.Rxg6+ Kxg6 2.Rh3 Bxg5 3.Qxh5+ Kf5 4.Rf3+ Ke5 5.Rxf7 Qh6 6.Qe2+ Be3 7.fxe3 >>>>> +- (7.53) Depth: 1/7 00:00:00 >>>>>1.Rxg6+ Kxg6 2.Rh3 Bxg5 3.Qxh5+ Kf5 4.Rf3+ Ke5 5.Rxf7 Qh6 6.Qe2+ Be3 7.fxe3 >>>>> +- (7.53) Depth: 1/7 00:00:00 >>>>>1.Rxg6+ Kxg6 2.Rh3 Bxg5 3.Qxh5+ Kf5 4.Rf3+ Ke5 5.Rxf7 Qh6 6.Qe2+ Be3 7.fxe3 >>>>> +- (7.53) Depth: 2/7 00:00:00 >>>>>1.Rxg6+ Kxg6 2.Rh3 Bxg5 3.Qxh5+ Kf5 4.Rf3+ Ke5 5.Rxf7 Qh6 6.Qe2+ Be3 7.fxe3 >>>>> +- (7.53) Depth: 3/10 00:00:00 >>>>>1.Rxg6+ Kxg6 2.Rh3 Bxg5 3.Qxh5+ Kf5 4.Rf3+ Ke5 5.Rxf7 Qh6 6.Qe2+ Be3 7.fxe3 >>>>> +- (7.53) Depth: 4/10 00:00:00 >>>>>1.Rxg6+ Kxg6 2.Rh3 Bxg5 3.Qxh5+ Kf5 4.Rf3+ Ke5 5.Rxf7 Qh6 6.Qe2+ Be3 7.fxe3 >>>>> +- (7.53) Depth: 5/11 00:00:00 >>>>>1.Rxg6+ Kxg6 2.Rh3 Bxg5 3.Qxh5+ Kf5 4.Rf3+ Ke5 5.Rxf7 Qh6 6.Qe2+ Be3 7.fxe3 >>>>> +- (7.53) Depth: 6/13 00:00:00 16kN >>>>>1.Rxg6+ Kxg6 2.Rh3 Bxg5 3.Qxh5+ Kf5 4.Rf3+ Ke5 5.Rxf7 Qh6 6.Qe2+ Be3 7.fxe3 >>>>> +- (7.53) Depth: 7/17 00:00:00 86kN >>>>>1.Rxg6+ Kxg6 2.Rh3 Bxg5 3.Qxh5+ Kf5 4.Rf3+ Ke5 5.Rxf7 Qh6 6.Qe2+ Be3 7.fxe3 >>>>> +- (7.53) Depth: 8/20 00:00:00 202kN >>>>>1.Rxg6+ Kxg6 2.Rh3 Bxg5 3.Qxh5+ Kf5 4.Rf3+ Ke5 5.Rxf7 Qh6 6.Qe2+ Be3 7.fxe3 >>>>> +- (7.53) Depth: 9/20 00:00:00 448kN >>>> >>>>I suspect that it learned from previous analysis. >>>>The line and the score was not changed from depth 1. >>>> >>>>Uri >>> >>> >>>Which previous analysis? I just copied the fen into Ruffian and let him >>>analyse. No previous analysis. Wrong suspicion :) Ruffian is a very smart >>>program... >>> >>>Djordje >> >>It does not make sense that a program has a pv of 13 plies at depth 1 when the >>pv and the score are unchanged until depth 9 except number of nodes. >> > >It is possible due to extensions. At depth 1 the PV is extended 13 plies deep. >The following iterations can produce the same pv and therefore the same score if >the extensions at depth 1 already produced the best sequence of moves up to ply >13. > >This behaviour happens also with a simple full-width search followed by a >capture search: if at the horizon of the full-width search - on the pv - a >capture is forced, the next iteration will have the same pv with the same depth >and score. > >Rgds, >Alessandro It is possible in theory but it does not make sense in this case when part of the moves are no captures or checks and the pv is not changed until depth 9. 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.