Author: Uri Blass
Date: 21:57:07 10/16/05
Go up one level in this thread
On October 16, 2005 at 19:26:59, Uri Blass wrote: >On October 16, 2005 at 18:56:56, Robert Hyatt wrote: > >>On October 15, 2005 at 16:40:42, Uri Blass wrote: >> >>>I played a blitz game. >>> >>>I did not play well in the opening and got the following position >>>It was obvioud for me that I have to play Ne8 >>> >>>surprisingly Fruit default never considers that move and prefer sacrificing a >>>piece for 2 pawns. >>> >>>I wonder if the sacrifice is good. >>>This sacrifice also change my opinion about fruit and now I am not sure if >>>increasing the value of pawns and pawn structure to 102 is really productive for >>>fruit. >> >>I believe this is the wrong way to tune any program. Otherwise you either see >>piece for two pawn sacs all the time, and they are generally bad, even if it is >>a piece for three pawns. Hence the "bad trade" code in Crafty, that simply >>directly addresses this problem to prevent it for all cases... unless the trade >>can lead to a draw... > >Interesting that fruit does not like to trade piece for 2 pawns all the >time(otherwise it could not be so strong). >In the game against Junior in WCCC it was Deep Junior that played the trade and >lost. > >Interesting to find now that surprisingly Fruit also likes the trade and has >positive score some moves later so maybe the error of Junior was later in the >game and the trade was not so bad(it may be interesting if programs can beat >Fruit with black after the trade). > >some analysis by Fruit(history threshold=50 that seems to be even stronger than >original fruit) > >59: Deep Junior - Fruit, 13th WCCC Reykjavik 2005 >[D]r2q1rk1/pp2npp1/1bn1b2p/3p2BQ/8/1NNB4/PPP2PPP/2KR3R w - - 0 1 > >Analysis by Fruit 2.2: > >13.Be3 Bxe3+ 14.fxe3 > ² (0.29) Depth: 1/11 00:00:00 >13.Be3 Bxe3+ 14.fxe3 Ng6 15.Bxg6 fxg6 16.Qxg6 > ² (0.32) Depth: 2/11 00:00:00 >13.Kb1 Bxf2 > ² (0.42) Depth: 2/11 00:00:00 >13.Kb1 Bxf2 14.Rhf1 > = (0.21) Depth: 3/11 00:00:00 >13.Kb1 Nb4 14.Be4 Bxf2 > ² (0.26) Depth: 4/12 00:00:00 >13.Kb1 Nb4 14.Be4 Bxf2 15.Nxd5 Nbxd5 16.Bxd5 Bxd5 17.Bxe7 Qxe7 18.Rxd5 > = (0.11) Depth: 5/13 00:00:00 >13.Kb1 Nb4 14.Qh4 Bxf2 15.Qxb4 hxg5 16.Qxb7 > ² (0.32) Depth: 6/23 00:00:01 73kN >13.Kb1 Nb4 14.Be4 Re8 15.Bh4 Qc7 16.Bg3 > ² (0.33) Depth: 7/23 00:00:01 166kN >13.Kb1 Qd7 14.Bxe7 Nxe7 15.Bb5 Qc8 16.Rd2 Bf5 17.Re1 > = (0.13) Depth: 8/24 00:00:01 552kN >13.Kb1 Qd7 14.Bxe7 Nxe7 15.Qh4 Rae8 16.Rhe1 Bg4 17.f3 Bf5 18.Bxf5 Nxf5 > = (0.23) Depth: 9/31 00:00:03 1603kN >13.Kb1 Qd7 14.Bxe7 Nxe7 15.Bb5 Qd6 16.Nd4 Nc6 17.Nxe6 fxe6 18.Ne4 Qf4 > = (0.14) Depth: 10/31 00:00:06 3362kN >13.Kb1 Qd7 14.Bxe7 Nxe7 15.Bb5 Qc7 16.Qf3 Qd6 17.h4 a6 18.Bd3 f6 > = (0.15) Depth: 11/39 00:00:14 8269kN >13.Kb1 Qd7 14.Bxe7 Nxe7 15.Qh4 Rfe8 16.Rhe1 Rac8 17.Nd4 Bg4 18.Be2 Bxd4 19.Rxd4 >Bxe2 20.Rxe2 > = (0.10) Depth: 12/42 00:00:29 17874kN >13.Bxh6 gxh6 14.Qxh6 Bf5 15.Bxf5 Nxf5 16.Qf4 Qf6 17.Rxd5 Nce7 18.g4 Nxd5 19.Nxd5 >Qh6 20.gxf5 Qxf4+ 21.Nxf4 Bxf2 > ² (0.39) Depth: 12/48 00:00:49 30956kN >13.Bxh6 gxh6 14.Qxh6 Bf5 15.Bxf5 Nxf5 16.Qf4 Qf6 17.Rxd5 Nce7 18.g4 Nxd5 19.Nxd5 >Qh6 20.gxf5 Qxf4+ 21.Nxf4 Bxf2 22.Kb1 > = (0.19) Depth: 13/50 00:01:13 46849kN >13.Bxh6 gxh6 14.Qxh6 Bf5 15.Bxf5 Nxf5 16.Qf4 Qf6 17.g4 Qh6 18.g5 Qh3 19.Nxd5 >Rfd8 20.Kb1 Rac8 21.Rhe1 > ² (0.39) Depth: 14/52 00:03:00 117073kN > >(, 17.10.2005) > > >Uri I continue to analyze and it seems based on fruit's analysis that Junior9 blundered in move 17 when it played 17.g4 Giving fruit with history threshold=50 to search 2 options it find the following analysis after finishing iteration 18. 59: Deep Junior - Fruit, 13th WCCC Reykjavik 2005 [D]r4rk1/pp3p2/1bn2q2/3p1n2/5Q2/1NN5/PPP2PPP/2KR3R w - - 0 1 Analysis by Fruit 2.2: 1. ² (0.40): 17.Rxd5 Nce7 18.Kb1 Rad8 19.Re5 Qg7 20.Ree1 Rd7 21.Ne4 Kh8 22.g4 Nd4 23.Qg3 Nd5 24.c4 f5 25.gxf5 Nxb3 26.Qh3+ Qh7 27.Qxh7+ Kxh7 28.cxd5 2. ² (0.34): 17.Kb1 Qe5 18.Qg4+ Ng7 19.Rxd5 Qe6 20.Qg3 Rfe8 21.Rg5 Qh6 22.Nd5 Rad8 23.h4 Re2 24.Nf4 Re7 25.h5 Kf8 26.Nd5 (, 17.10.2005) 17.g4 leads to negative score for white based on analysis with fruit. Fruit2.2(hostory=50) can see negative score for white after 17.g4 Qh6 18.g5(forced) Qg6 and here is the analysis 59: Deep Junior - Fruit, 13th WCCC Reykjavik 2005 [D]r4rk1/pp3p2/1bn3q1/3p1nP1/5Q2/1NN5/PPP2P1P/2KR3R w - - 0 1 Analysis by Fruit 2.2: 19.Nxd5 = (0.02) Depth: 1/1 00:00:00 19.Nxd5 Rfd8 = (0.15) Depth: 2/11 00:00:00 19.Rxd5 Ncd4 20.Nxd4 Nxd4 21.Rxd4 Bxd4 22.Qxd4 Qxg5+ 23.Kb1 ² (0.60) Depth: 2/11 00:00:00 19.Rxd5 Nce7 20.Rd7 ² (0.45) Depth: 3/11 00:00:00 19.Rxd5 Ncd4 20.Nxd4 Nxd4 21.Rxd4 Bxd4 22.Qxd4 Qxg5+ 23.Kb1 ² (0.60) Depth: 4/15 00:00:00 19.Rxd5 Nfe7 20.Rd7 Qf5 21.Qd2 Qxf2 22.Rxb7 Qxd2+ 23.Kxd2 = (0.21) Depth: 5/16 00:00:00 19.Nxd5 Rfd8 20.Kb1 Bd4 21.Nxd4 Rxd5 ² (0.42) Depth: 5/16 00:00:00 19.Nxd5 Rfd8 20.h4 Rac8 21.h5 Qe6 ² (0.50) Depth: 6/18 00:00:01 44kN 19.Nxd5 Rfd8 20.h4 Rac8 21.Kb1 Ncd4 22.Nxd4 Nxd4 23.Rxd4 Bxd4 24.Qxd4 Rxc2 ² (0.45) Depth: 7/18 00:00:01 95kN 19.Nxd5 Rad8 20.h4 Nfd4 21.Nxd4 Rxd5 22.Nxc6 Qxc6 23.Rxd5 Qxd5 ² (0.48) Depth: 8/22 00:00:01 295kN 19.Nxd5 Rad8 20.Nxb6 axb6 21.h4 Qe6 22.Rde1 Qd5 23.Kb1 Rfe8 ² (0.47) Depth: 9/22 00:00:01 578kN 19.Nxd5 Rad8 20.Nxb6 axb6 21.h4 Nfe7 22.Kb1 Nd5 23.Qc4 Ne5 24.Qb5 ² (0.35) Depth: 10/26 00:00:03 1315kN 19.Nxd5 Rfd8 20.Rhe1 Bd4 21.Nxd4 Rxd5 22.Nxc6 Qxc6 23.Rxd5 Qxd5 24.Rd1 Qe6 25.h4 ² (0.40) Depth: 11/36 00:00:07 4612kN 19.Nxd5 Rfd8 20.Rhe1 Nfd4 21.Nxd4 Rxd5 22.Nxc6 Qxc6 23.Rxd5 Qxd5 24.c4 Qc5 25.Re3 Rd8 ² (0.29) Depth: 12/36 00:00:16 11251kN 19.Nxd5 Rfd8 20.Rhe1 Nfd4 21.Nxd4 Rxd5 22.Nxc6 Qxc6 23.Rxd5 Qxd5 24.Kb1 Qg2 25.f3 Rc8 26.c3 = (0.16) Depth: 13/36 00:00:35 25149kN 19.Nxd5 Rfd8 20.Kb1 Bd4 21.Nxd4 Rxd5 22.Nxc6 Qxc6 23.Rhe1 Rad8 24.Rxd5 Qxd5 25.b3 Qa5 26.b4 Qb5 = (0.00) Depth: 14/48 00:01:54 82533kN 19.h4 Nfe7 20.h5 Qe6 21.Kb1 Rae8 22.a3 Qf5 23.Qxf5 Nxf5 24.Rxd5 Nfd4 25.Na4 Nxb3 26.Nxb6 = (0.16) Depth: 14/48 00:02:26 105313kN 19.h4 Nfe7 20.h5 Qe6 21.Kb1 d4 22.Ne4 Nd5 23.Qh4 Rfe8 24.Rde1 Ne5 25.h6 Rac8 26.f4 = (0.19) Depth: 15/48 00:03:13 138671kN 19.h4 d4 20.h5 Qe6 21.Rhe1 Ne3 22.Ne4 Ng2 23.Nf6+ Kh8 24.Qe4 Qxe4 25.Rxe4 Bd8 26.Nxd4 Bxf6 27.gxf6 Nxd4 28.Rdxd4 Rg8 ² (0.34) Depth: 16/48 00:07:17 312368kN 19.h4 Nfe7 20.h5 Qe6 21.Kb1 Rad8 22.Rhe1 Qf5 23.Qxf5 Nxf5 24.Rxd5 Bxf2 25.Rf1 Ne3 26.Rxd8 Nxf1 27.Rd7 f6 = (0.21) Depth: 17/53 00:12:08 519660kN 19.h4 Nfe7 20.h5 Qe6 21.Kb1 Rad8 22.Rhf1 Rfe8 23.Rfe1 Qf5 24.Qxf5 Nxf5 25.Rxe8+ Rxe8 26.Nxd5 Bd8 27.f4 Re2 28.Nc5 b6 29.Nc3 = (0.12) Depth: 18/58 00:23:09 976951kN 19.h4 Nfe7 20.h5 Qh7 21.Kb1 Rad8 22.a3 Rfe8 23.Qf3 Ne5 24.Qf4 Nc4 25.h6 Rc8 26.Qf3 Qf5 27.h7+ Kg7 28.Qxf5 Nxf5 29.Rxd5 = (-0.03) Depth: 19/58 01:06:33 2773733kN (, 17.10.2005) iteration 19 was finished and fruit could see nothing better than 19.h4 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.