Author: Uri Blass
Date: 06:22:43 11/20/02
Go up one level in this thread
On November 20, 2002 at 08:54:34, Uri Blass wrote: >On November 20, 2002 at 08:14:52, Odd Gunnar Malin wrote: > >>On November 20, 2002 at 07:40:57, Darren Rushton wrote: >> >>>What do most people consider a reasonable resign score to set in a .ini file? >>> >>>I looked at a few GM games and set Fritz, Hiarcs, whoever to analyse the score >>>when resignation occured. >>> >>>I concluded that between 400 and 500 centipawns is about the point, although >>>most programmers set to 700 as a default. >>> >> >>Just a little remark. >> >>I had the following line in my book: >>1.e4 e5 2.Nf3 Nc6 3.Bc4 Nf6 4.Ng5 Bc5 5.Bxf7 >> >>An exotic line in a book build upon some GM games. All GM's played 5.Bxf7 here >>as White but when I lunched the engine on ICC I saw that most players would play >>5.Nxf7 instead. >>Check your engine score here and after the right 5...Bxf2! and see if it will >>resign if it play as Black. >> >>The statistics say 73% score for Black! >> >>If you also use movecount in the resign algorithm you could adjust the resign >>score based upon opponent rating instead of having it constant. >>There is two factors to consider: You don't want to resign to fast, and you >>don't want to bother your opponent so that he don't comes back to play against >>your engine. >> >>Odd Gunnar > >Analysis by movei after Bxf2+ > >depth=11 +3.45 e1f1 d8e7 f7h8 f2h4 d2d3 d7d5 e4d5 e7c5 d1e2 c8g4 e2e3 >Nodes: 8164098 NPS: 150656 >Time: 00:00:54.19 > >depth=11 +3.45 e1f1 d8e7 f7h8 f2h4 d2d3 d7d5 e4d5 e7c5 d1e2 c8g4 e2e3 >Nodes: 17416198 NPS: 154467 >Time: 00:01:52.75 > >depth=12 +3.48 e1f1 d8e7 f7h8 f2b6 d2d3 d7d5 e4d5 e7c5 d1f3 c8g4 f3g3 c6e7 >Nodes: 27342733 NPS: 156539 >Time: 00:02:54.67 > >depth=12 +3.48 e1f1 d8e7 f7h8 f2b6 d2d3 d7d5 e4d5 e7c5 d1f3 c8g4 f3g3 c6e7 >Nodes: 35572204 NPS: 155486 >Time: 00:03:48.78 > >depth=13 +3.24 e1f1 d8e7 f7h8 f2b6 d2d3 e7c5 d1e1 d7d5 e4d5 f6d5 c4d5 c5d5 b1c3 >d5d4 >Nodes: 61369731 NPS: 156316 >Time: 00:06:32.60 > >I do not plan to resign at -4 but even with resigning at -4 movei is not going >to resign as black here. > >Note that movei does not have king safety evaluation and I can imagine that with >king safety evaluation the score is going to be even less optimistic for white. > >Uri scores continues to go down and I stop the analysis depth=14 +3.09 e1f1 d8e7 f7h8 f2b6 d2d3 e7c5 d1e1 d7d5 e4d5 f6d5 e1h4 c8e6 b1c3 c5f8 f1e1 Nodes: 262355892 NPS: 154555 Time: 00:28:17.49 After Bxf7+ Movei can see only a small advantage for white. depth=11 -0.57 e8e7 f7d5 c6b4 b1c3 h8f8 d5b3 d7d6 g5f3 c8g4 d2d3 e7e8 Nodes: 11209785 NPS: 153685 Time: 00:01:12.94 depth=12 -0.70 e8e7 f7c4 h8f8 b1c3 c5f2 e1f2 f6e4 f2g1 e4g5 d2d4 e7e8 d1h5 g5f7 c4f7 f8f7 d4e5 d7d6 e5d6 d8d6 h5h7 Nodes: 46251526 NPS: 154171 Time: 00:05:00.00 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.