Author: chandler yergin
Date: 18:37:32 12/01/05
Go up one level in this thread
On December 01, 2005 at 19:54:17, Uri Blass wrote: >On December 01, 2005 at 19:45:01, George Tsavdaris wrote: > >>>Look at the 'bottom' line. >>> = (0.05) Depth: 17/51 00:03:21 165911kN >>> = (-0.17) Depth: 17/69 00:03:39 183051kN >>> >>>The depth of search of 'interesting' variations is significant. >>>Also nodes searched. The analysis should not be the same. >>> >> Actually the common logic says that if you analyse with engine-A a position-B >>no matter how this position occured in the board for analysis, if you pasted the >>FEN or went to it from a PGN notation, the engine-A has to show the same >>analysis each time........(I assume that hash tables were cleared before >>analysing the position for the 2nd time) > >Yes >I replaced engines and cleared the hash. >The history may be important in case of repetition that may be evaluated as a >draw in one of the cases but here the last move was a pawn move so this is not >the reason. > > >Note that the bigger difference in the analysis is in depth 15 both the move and >the evaluation are different. You're doing the same thing again, ignoring the Depth of the Variations and nodes searched. You can't do that and claim it's a bug. The Engine is doing what engines do. ³ (-0.49) Depth: 15/47 00:00:49 39631kN = (0.07) Depth: 15/51 00:00:54 42522kN > > >1.gxh5 c2+ 2.Kc1 Rxe5 3.Rxe5 fxe5 4.Rd3 gxh5 5.Qxh5 Rxd6 6.Qxe5 Qh1+ 7.Kxc2 Qg2+ >8.Kc3 Qc6+ 9.Kd2 Qg2+ 10.Qe2 Qg5+ 11.Kc2 Rc6+ 12.Rc3 Qg6+ 13.Kc1 Qg1+ 14.Kc2 > = (0.07) Depth: 15/51 00:00:54 42522kN > >34.bxc3 Rxe5 35.gxh5 Rxd5 36.Rxd5 Qxd5 37.Qxd7 Qb3+ 38.Kc1 Qxc3+ 39.Kd1 Qd3+ >40.Ke1 Qe3+ 41.Kf1 Qc1+ 42.Ke2 Qc4+ 43.Ke1 Qh4+ 44.Kf1 Qf4+ 45.Kg2 Qd2+ 46.Kf3 > ³ (-0.49) Depth: 15/47 00:00:49 39631kN > >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.