Test position

Discussion of anything and everything relating to chess playing software and machines.

Moderator: Ras

kgburcham
Posts: 2016
Joined: Sun Feb 17, 2008 4:19 pm

Test position

Post by kgburcham »

The top commercial programs find this quick. What about the other programs.

[d] 2r4k/1p1b2pp/p6q/5r1n/3P2QP/P1N2pP1/BP3P2/2R2RK1 b - - Nxg3

Little Thought v1.03 mp 64

13/32 0:16 -0.97 28...Nf6 29.Qh3 Qd2 30.Rb1 Rb5 31.Be6 Rxb2 32.Bxd7 Rxc3 33.Qe6 Rbc2 34.Rb4 (18.430.473) 1086
14/40 0:56 -1.17++ 28...Nxg3 29.Rfd1 Ne4 30.Qxe4 Rf4 31.Qxb7 Qxh4 32.Be6 Bxe6 33.Qxg7+ Kxg7 34.Kf1 Rxd4 35.Rxd4 Qxd4 36.Re1 Qxf2+ 37.Kxf2 Bg4 38.Re7+ Kh6 (47.622.479) 843
15/40 1:32 -1.43++ 28...Nxg3 29.Qxg3 Rf4 30.Bd5 Rg4 31.Bxf3 Rxg3+ 32.fxg3 Bh3 33.Kh2 Bxf1 34.Kh1 (85.294.271) 923
15/42 2:00 -1.58 28...Nxg3 29.Qxg3 Rf4 30.Bd5 Rg4 31.Bxf3 Rxg3+ 32.fxg3 Bh3 33.Bg2 Bxg2 34.Kxg2 Qd2+ 35.Kh3 Qxb2 36.Na4 Qxc1 37.Rxc1 Rxc1 38.Nc5 (117.187.165) 975
best move: Nh5xg3 time: 2:43.156 min n/s: 1.055.409 nodes: 150.000.008
User avatar
pedrox
Posts: 1056
Joined: Fri Mar 10, 2006 6:07 am
Location: Basque Country (Spain)

Re: Test position

Post by pedrox »

4 s for danasah

FEN: 2r4k/1p1b2pp/p6q/5r1n/3P2QP/P1N2pP1/BP3P2/2R2RK1 b - - 0 1

Danasah363frcpgo:
2 00:00 1.368 1.368 -0,56 Qd2 Ne4 Qxb2 Rxc8+ Bxc8
3 00:00 2.071 2.071 -0,56 Qd2 Ne4 Qxb2 Rxc8+ Bxc8
4 00:00 4.165 416.500 -0,56 Qd2 Ne4 Qxb2 Rxc8+ Bxc8
4 00:00 7.041 704.100 +0,04 Nf6 Qh3 Rd5 Qh1
5 00:00 16.294 543.133 -0,46 Nf6 Qh3 Rd5 Qh1 Rf5
6 00:00 28.822 720.550 -0,41 Nf6 Qh3 Qd2 g4 Rb5 Qxf3 Nxg4
7 00:00 61.547 879.242 +0,26 Nf6 Qh3 Qd2 Rb1 Qxd4 Rbd1 Qc5
8 00:00 100.794 916.309 +0,22 Nf6 Qh3 Qd2 Rb1 Rb5 Be6 Rxb2 Rxb2 Qxb2
9 00:00 196.149 980.745 +0,44 Nf6 Qh3 Qd2 Rb1 Rb5 Be6 Rxb2 Rxb2 Qxb2 Bxd7 Rxc3
10 00:00 365.553 1.015.425 +0,44 Nf6 Qh3 Qd2 Rb1 Rb5 Be6 Rxb2 Rxb2 Qxb2 Bxd7 Rxc3
11 00:00 663.778 1.070.609 +0,51 Nf6 Qh3 Qd2 Rb1 Rb5 Be6 Rxb2 Rxb2 Qxb2 Bxd7 Rxc3 a4
12 00:01 1.603.351 1.068.900 +0,80 Nf6 Qh3 Qd2 Rb1 Rb5 Be6 Rxb2 Rxb2 Qxb2 Ne4 Nxe4 Bxd7 Rc2
12 00:04 4.183.268 981.987 +1,54 Nxg3 Qxg3 Rf4 Bd5 Rg4 Bxf3 Rxg3+ fxg3 Qe3+ Kh1 Rb8 Ne2 Rf8 Rc3
13 00:06 5.914.068 984.037 +1,40 Nxg3 Qxg3 Rf4 Bd5 Rg4 Bxf3 Rxg3+ fxg3 Qe3+ Kh1 b6 Rcd1 Bh3 Nd5 Qh6
14 00:08 7.984.521 966.649 +1,32 Nxg3 Qxg3 Rf4 Bd5 Rg4 Bxf3 Rxg3+ fxg3 Qe3+ Kh1 b6 Rcd1 Bh3 Nd5 Qh6 Rfe1
15 00:17 16.398.937 964.076 +1,57 Nxg3 Qxg3 Rf4 Bd5 Rg4 Bxf3 Rxg3+ fxg3 Qe3+ Kh1 b6 d5 Qe5 Kg2 Kg8 Rce1 Qd4
16 00:35 33.130.682 946.590 +1,77 Nxg3 Qxg3 Rf4 Bd5 Rg4 Bxf3 Rxg3+ fxg3 Qe3+ Kh1 b6 d5 Qd2 Ne4 Qxb2 Rxc8+ Bxc8 Nd6 Bd7
User avatar
Ovyron
Posts: 4562
Joined: Tue Jul 03, 2007 4:30 am

Re: Test position

Post by Ovyron »

Alaric instantly:

CPU0: AuthenticAMD x86 Family 6 Model 10 Stepping 0 2133 MHz
Engine: Alaric 707 (64 MB)
by Peter Fendrich
9.00 0:00 -0.40 1...Nf6 2.Qh3 Qd2 3.Rb1 Rb5 4.Be6 Rxb2 5.Rxb2 Qxb2 (137.132)
9.00 0:00 -0.82 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 Bb5 (436.390)
10.00 0:01 -0.80 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 Qxd4 7.Bxb7 (584.441) 561
11.01 0:01 -0.95 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Bh3 6.Rfd1 Qe3+ 7.Kh2 Qxf3 8.Kxh3 (897.890) 553
12.01 0:03 -1.52 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 Bc6 7.Bxc6 bxc6 8.Kh2 Qd2+ 9.Kh3 Qxb2 (1.885.851) 570
13.01 0:05 -1.65 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qd2 6.Bxb7 Bh3 7.Na2 Rxc1 8.Rxc1 Qxd4+ 9.Kh2 Qxb2+ 10.Kxh3 Qxb7 (3.421.620) 573
Break set from 'stop'
best move: Nh5xg3 time: 0:09.093 min n/s: 572.336 nodes: 5.049.717

Last Movei with the 101010 settings on 1 sec, but then it loses it and finds it again after 11 secs, I wonder what time would be recorded?

[Snip]
9.00 0:00 -0.79 1...Nf6 2.Qh3 Qd2 3.Rb1 Qxd4 4.Rbd1 Re5 5.Rxd4 Bxh3 (157.061)
9.45 0:00 -0.79 1...Nf6 2.Qh3 Qd2 3.Rb1 Qxd4 4.Rbd1 Re5 5.Rxd4 Bxh3 (300.765)
10.01 0:00 -1.00 1...Nf6 2.Qh3 Qd2 3.Rb1 Re5 4.Qh1 Qxd4 5.Rfd1 Qg4 6.Rbc1 (362.451)

10.02 0:01 -1.01 1...Nxg3 (539.192)
10.02 0:01 -1.30 1...Nxg3 (711.042)
10.02 0:03 -1.32 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kg2 Qd2+ 7.Ne2 Rf8 (1.179.405)
10.45 0:03 -1.32 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kg2 Qd2+ 7.Ne2 Rf8 (1.278.044)
11.01 0:04 -1.09 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 Bh3 7.Nd5 Qxd4 8.Rxc8+ Bxc8 (1.514.179)

11.02 0:04 -1.10 1...Nf6 (1.585.307)
11.02 0:04 -1.10 1...Nf6 2.Qh3 Qd2 3.Rb1 Rb5 4.Be6 Rxb2 5.Bxd7 Rxc3 6.Rxb2 Qxb2 (1.692.760)
11.45 0:08 -1.10 1...Nf6 2.Qh3 Qd2 3.Rb1 Rb5 4.Be6 Rxb2 5.Bxd7 Rxc3 6.Rxb2 Qxb2 (3.132.997)
12.01 0:09 -1.30 1...Nf6 2.Qh3 Qd2 3.Rb1 Rb5 4.Be6 Rxb2 5.Bxd7 Rxc3 6.Rxb2 Qxb2 7.a4 Rc4 (3.585.692)

12.02 0:11 -1.31 1...Nxg3 (4.062.497)
12.02 0:12 -1.40 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Bh3 6.Rfe1 Qd2 7.Bxb7 Qxb2 8.Bc6 (4.524.751)
12.45 0:14 -1.40 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Bh3 6.Rfe1 Qd2 7.Bxb7 Qxb2 8.Bc6 (5.304.604)
13.01 0:18 -1.39 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Bh3 6.Rfe1 Qd2 7.Bxb7 Qxb2 8.Bd5 Rf8 (6.519.649)
13.45 0:23 -1.39 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Bh3 6.Rfe1 Qd2 7.Bxb7 Qxb2 8.Bd5 Rf8 (8.422.928)


Glaurung 2.0.1 on 1 sec:

[Snip]
9.00 0:00 -0.05 1...Nf6 2.Qh3 Qd2 3.Rb1 Qxd4 4.Rbd1 Re5 5.Rxd4 Bxh3 (66.994) 163
10.00 0:00 +0.01 1...Nf6 2.Qh3 Qd2 3.Rb1 Qxd4 4.Rbd1 Re5 5.Rxd4 Bxh3 6.Rfd1 (115.418) 235
11.00 0:00 -0.07 1...Nf6 2.Qh3 Qd2 3.Rb1 Rxc3 4.Rfd1 Qe2 5.bxc3 Rb5 6.Qh1 Qxa2 7.Rxb5 Bxb5 8.Qxf3 Qxa3 (279.740) 372
11.00 0:01 -0.62 1...Nxg3 2.Rfd1 Nh5 3.Kf1 Qf4 4.Qxf4 Nxf4 5.Bb1 Rh5 6.Be4 Rxh4 7.Bxf3 (1.015.582) 525
12.01 0:03 -1.19 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kg2 Qd2+ 7.Ne2 Re8 8.Rc7 Rxe2+ 9.Bxe2 Qxe2+ 10.Rf2 (1.869.386) 563
13.01 0:04 -1.13 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 Bh3 7.Bxb7 Bxf1 8.Rxf1 Re8 (2.376.373) 583
14.01 0:06 -1.37 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 Rb8 7.Kg2 Qd2+ 8.Ne2 Bb5 9.Rfe1 Qxb2 (3.874.910) 605
best move: Nh5xg3 time: 0:08.733 min n/s: 619.345 nodes: 5.340.000

Toga 3.1.2SE JA Chekov instantly:

[Snip]
7/20 0:00 -1.01 1...Nf6 2.Qh3 Qd2 3.g4 Rf4 4.Kh1 Rxg4 (16.530)
8/22 0:00 -0.63 1...Nf6 2.Qh3 Qd2 3.Rb1 Rb5 4.Be6 Rxb2 5.Rxb2 Qxb2 6.Bxd7 Rxc3 (64.766)
9/25 0:00 -0.63 1...Nf6 2.Qh3 Qd2 3.Rb1 Rb5 4.Be6 Rxb2 5.Rxb2 Qxb2 6.Bxd7 Rxc3 (164.549)
9/36 0:00 -1.03++ 1...Nxg3 2.fxg3 Qd2 3.Qh3 Qxb2 4.g4 Rf4 5.Qg3 Rxd4 6.Qh2 Rxg4+ 7.Kh1 (568.212)
10/43 0:03 -1.48++ 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 Bh3 7.Bxb7 Bxf1 8.Rxf1 Rb8 9.Bxa6 Qxg3 (2.511.974) 655
11/43 0:05 -1.46 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 Rd8 7.Rce1 Qxd4 8.Bxb7 Qb6 9.Bd5 Qxb2 (3.374.816) 665
12/43 0:07 -1.51 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 Rd8 7.Rce1 Qxd4 8.Bd5 h5 9.Re7 Bb5 10.Rf4 (5.320.919) 668
best move: Nh5xg3 time: 0:10.365 min n/s: 658.789 CPU 100.0% n/s(1CPU): 658.789 nodes: 6.690.000
Harald Johnsen

Re: Test position

Post by Harald Johnsen »

11 sec for Cyrano 0.4 on a slow p4.

10 64 1.53 344742 Nf6 Qh3 Qd2 Rb1 Rxc3 bxc3 Rd5 Qh1 Qxa2 Rxb7 Qe2 Rb8+ Be8
11 58 3.09 731388 Nf6 Qh3 Qd2 Rb1 Rxc3 bxc3 Rd5 Rxb7 Qxa2 Rb8+ Ng8 g4 Qe2
11 59 3.85 940176 Nxg3 Rfd1
++
11 128 10.75 2810644 Nxg3 Qxg3 Rf4 Bd5 Rg4 Bxf3 Rxg3+ fxg3 Qe3+ Kh1 Bh3 Bxb7 Bxf1 Rxf1 Rd8 Bxa6 Rxd4
12 136 14.39 3784003 Nxg3 Qxg3 Rf4 Bd5 Rg4 Bxf3 Rxg3+ fxg3 Qe3+ Kh1 Bh3 Bxb7 Bxf1 Rxf1 Rb8 Bxa6 Rxb2


HJ.
royb
Posts: 577
Joined: Thu Mar 09, 2006 12:53 am

Re: Test position

Post by royb »

I realize you wanted to see how non-commercial programs do with this position, but I tried Rybka 2.3.2a 32bit on this and it seems that there may be a stronger move than ...Nxg3 here:

Analysis by Rybka 2.3.2a 32-bit :

1...Nf6 2.Qh3
µ (-0.71) Depth: 5 00:00:00
1...Nf6 2.Qh3 Qd2
³ (-0.61) Depth: 6 00:00:00 2kN
1...Nf6 2.Qh3 Qd2 3.Bb1
³ (-0.70) Depth: 7 00:00:00 4kN
1...Nf6 2.Qh3 Qd2 3.Bb1 Rg5
µ (-0.80) Depth: 8 00:00:00 7kN
1...Nf6 2.Qh3 Qd2 3.Bb1 Rg5 4.Rc2
µ (-0.74) Depth: 9 00:00:00 11kN
1...Nf6 2.Qh3 Qd2 3.Bb1 Rh5 4.Qh1 Bg4
³ (-0.57) Depth: 10 00:00:00 30kN
1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 Bh3
µ (-0.95) Depth: 10 00:00:01 97kN
1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 b6 7.Rce1
µ (-0.97) Depth: 11 00:00:02 165kN
1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 b6 7.Rce1 Qxd4 8.Bb7
µ (-0.81) Depth: 12 00:00:03 230kN
1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qe3+ 6.Kh1 b5 7.Rcd1 Rf8 8.Kg2
µ (-0.86) Depth: 13 00:00:06 410kN
1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qd2 6.Bxb7 Bh3 7.Rfe1 Qxb2 8.Bf3
µ (-1.17) Depth: 14 00:00:11 784kN
1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qd2 6.Rb1 Qxd4+ 7.Rf2 Qe5 8.Kg2
µ (-1.18) Depth: 15 00:00:17 1262kN
1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qd2 6.Rb1 Qxd4+ 7.Rf2 Qe5 8.Kg2
µ (-1.17) Depth: 16 00:00:29 2034kN
1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+ 5.fxg3 Qd2 6.Rb1 Qxd4+ 7.Rf2 Qd6 8.Kh2
µ (-1.22) Depth: 17 00:01:02 4224kN
1...Nf4 2.Kh2 Rh5 3.Qxf3 Rxh4+ 4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Rfd1 Rf6 7.Ne4 Nh3+ 8.Qxh3
-+ (-1.86) Depth: 17 00:01:30 6177kN
1...Nf4 2.Kh2 Rh5 3.Qxf3 Rxh4+ 4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Rfd1 Rf6 7.Ne4 Nh3+ 8.Qxh3
-+ (-1.93) Depth: 18 00:02:06 8529kN
1...Nf4 2.Kh2 Rh5 3.Qxf3 Rxh4+ 4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Rfd1 Rf6 7.Ne4 Nh3+ 8.Qxh3
-+ (-1.87) Depth: 19 00:03:02 11915kN
1...Nf4 2.Kh2 Rh5 3.Qxf3 Rxh4+ 4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Rfd1 Nh3+ 7.Kf1 Rf6 8.Qxf6
-+ (-2.29) Depth: 20 00:05:31 21747kN
1...Nf4 2.Kh2 Rh5 3.Qxf3 Rxh4+ 4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Rfd1 Nh3+ 7.Kf1 Rf6 8.Rd3
-+ (-2.29) Depth: 21 00:08:36 33890kN

Rybka gives a significantly higher score to ...Nf4 Is she right?

Roy
User avatar
Eelco de Groot
Posts: 4702
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Glaurung Mjolnir Re: Test position

Post by Eelco de Groot »

royb wrote: 1...Nf4 2.Kh2 Rh5 3.Qxf3 Rxh4+ 4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Rfd1 Nh3+ 7.Kf1 Rf6 8.Rd3
-+ (-2.29) Depth: 21 00:08:36 33890kN

Rybka gives a significantly higher score to ...Nf4 Is she right?

Roy
Yes! Glaurung 080420 Mjolnir can confirm! A beautiful move!

In spite of the fact that the King Safety is still computed with differences between White and Black just as described in an earlier post, the evaluation seems in reasonable balance.


2r4k/1p1b2pp/p6q/5r1n/3P2QP/P1N2pP1/BP3P2/2R2RK1 b - -

Engine: Glaurung 080420 Mjolnir 001 Build 5 (256 MB, Athlon 2009 MHz, single core 32 bits version of Glaurung)
by Tord Romstad

2.00 0:00 -0.62 1...Qd2 2.Ne4 Qxb2 3.Rxc8+ Bxc8 (270) 1

3.00 0:00 -0.62 1...Qd2 2.Ne4 Qxb2 3.Rxc8+ Bxc8 (708) 3

3.00 0:00 -0.50 1...Nf6 2.Qh3 Qd2 (978) 4

4.00 0:00 -0.13 1...Nf6 2.Qh3 Qd2 3.g4 (1.580) 7

5.00 0:00 -0.41 1...Nf6 2.Qh3 Qd2 3.g4 Rf4 (3.036) 12

6.00 0:00 +0.78 1...Nf6 2.Qh3 Qd2 3.Kh1 Qxb2 4.g4 (7.315) 31

7.00 0:00 +1.07 1...Nf6 2.Qh3 Qd2 3.Kh1 Qxb2 4.Bb1 Re5 (15.650) 59

8.00 0:00 +0.50 1...Nf6 2.Qh3 Qd2 3.Rb1 Qxd4 4.Rbd1 Qc5
5.Kh2 (39.112) 125

9.00 0:00 +0.49 1...Nf6 2.Qh3 Qd2 3.Rb1 Qxd4 4.Rbd1 Qc5
5.Kh1 Bc6 (63.100) 168

10.00 0:00 +0.58 1...Nf6 2.Qh3 Qd2 3.Rb1 Rb5 4.Be6 Rxb2
5.Rxb2 Qxb2 6.Bxd7 Rxc3 (112.691) 240

11.00 0:00 +0.78 1...Nf6 2.Qh3 Qd2 3.Rb1 Rb5 4.Be6 Rxb2
5.Rxb2 Qxb2 6.Nd5 Nxd5 7.Bxd7 (218.178) 317

12.01 0:02 +0.64 1...Nf6 2.Qh3 Qd2 3.Rb1 Rb5 4.Be6 Rxb2
5.Bxd7 Rxc3 6.Rbd1 Qc2 7.Kh2 Rxa3 (1.196.245) 447

12.02 0:04 +1.90 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qe3+ 6.Kh2 Qd2+ 7.Ne2 Re8
8.Kg2 Rf8 (2.195.204) 473

13.01 0:05 +1.90 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Bh3 7.Na2 Re8
8.Bc6 Qxd4+ 9.Kh1 (2.774.970) 483

14.01 0:09 +1.90 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Bh3 7.Na2 Re8
8.Bc6 Qxd4+ 9.Kh1 (4.528.452) 500

15.01 0:16 +1.78 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Bh3 7.Nb1 Qxd4+
8.Kh2 Qxb2+ 9.Kxh3 Qxb7 10.Rxc8+ Qxc8+
11.g4 (8.649.254) 514

16.01 0:35 +1.94 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Bh3 7.Nb1 Qxd4+
8.Kh2 Qxb2+ 9.Kxh3 Qxb7 10.Rxc8+ Qxc8+
11.Kg2 Qc6+ 12.Kf2 (18.707.356) 524

17.01 1:37 +1.92 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Re8 7.Bg2 Qxb2
8.Kh2 Qxa3 9.Rc2 Re3 10.Ne4 (51.403.475) 528

18.01 4:08 +1.96 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Re8 7.Bg2 Qxb2
8.Kh2 Qxa3 9.d5 Qd6 10.Rce1 Rb8 (133.235.543) 537


18.02 5:31 +3.56 1...Nf4 2.Kh1 Rh5 3.Qxf3 Rxh4+
4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Qg3 Rg6
7.Qxg6 hxg6 8.Bd5 Qg4+ 9.Kh2 Nxd5
10.Nxd5 Qh5+ 11.Kg1 Qxd5 12.Rc7 Kh7 (179.947.950) 542

19.01 8:09 +3.74 1...Nf4 2.Kh1 Rh5 3.Qxf3 Rxh4+
4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Qg3 Rg6
7.Bd5 Nxd5 8.Qxg6 hxg6 9.Nxd5 Qg4+
10.Kh2 Qh5+ 11.Kg3 Qxd5 12.Rfd1 Qb3+
13.Kf4 Qxb2 (266.938.218) 545

20.01 13:37 +3.60 1...Nf4 2.Kh1 Rh5 3.Qxf3 Rxh4+
4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Qg3 Rg6
7.Rc2 Rxg3+ 8.fxg3 Qxg3+ 9.Kh1 Qh3+
10.Kg1 Qe3+ 11.Kh1 g5 12.Rxf4 gxf4
13.Bd5 Bc6 14.Bxc6 bxc6 (448.190.803) 548

21.01 26:07 +3.78 1...Nf4 2.Kh1 Rh5 3.Qxf3 Rxh4+
4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Qg3 Rg6
7.Rc2 Rxg3+ 8.fxg3 Qxg3+ 9.Kh1 Qh3+
10.Kg1 g5 11.d5 h5 12.Bc4 Kg7 13.Rh2 Qe3+
14.Kh1 Bh3 (866.868.803) 553
User avatar
Eelco de Groot
Posts: 4702
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Re: 080420? Hopelessly out of date: Here is 080422!

Post by Eelco de Groot »

And this is with the latest version from today that Tord sent me. At first I could not compile it but Tord showed me how to do it. There are many new options. Maybe it is alright to quote Tord a little about what he wrote the last and previous version, because I think it is really interesting for everybody who would like to know a little more about the new Glaurung:

Tord wrote:
My test results with the previous version I sent to you were, as always, so confusing and contradictory that they are probably worthless. Glaurung won 54.5-45.5 against Hiarcs 11.1, which is very good, won by 60-40 against Fruit 2.3.1, which is phenomenal, and lost by 38.5-61.5 against Strelka 2.0B, which is catastrophic.

I now have the source code of a new version available for download. There are lots of new UCI parameters to play with; the most important ones are probably "King Safety Coefficient", "King Safety X Intercept", "King Safety Max Slope", "Queen Contact Check Bonus", "Queen Check Bonus", "Rook Check Bonus", "Bishop Check Bonus" and "Knight Check Bonus".

I don't have time right now to explain exactly what all these parameters do, but "King Safety Coefficient", "King Safety X Intercept" and "King Safety Max Slope" are used to compute the contents of the SafetyTable[] array (which was just a constant array in the previous version). The various "check bonuses" should be easy to understand if you look at the code. Basically I just count the number of safe checks for each piece type, multiply the number by the corresponding bonus, and add it to the 'inc' variable.
Thanks Tord!


[d]2r4k/1p1b2pp/p6q/5r1n/3P2QP/P1N2pP1/BP3P2/2R2RK1 b - -

Engine: Glaurung 080422 Mjolnir 001 Build 1 (256 MB)
by Tord Romstad

2.00 0:00 -0.82 1...Qd2 2.Kh1 Qxb2 (275) 0

3.00 0:00 -0.82 1...Qd2 2.Kh1 Qxb2 (718) 2

3.00 0:00 +0.07 1...Nf6 2.Qh3 Qd2 (972) 2

4.00 0:00 +0.35 1...Nf6 2.Qh3 Qd2 3.Kh2 Qxd4 (1.504) 4

5.00 0:00 +0.19 1...Nf6 2.Qh3 Qd2 3.g4 Rf4 (3.808) 11

6.00 0:00 +1.39 1...Nf6 2.Qh3 Qd2 3.Bb1 Rf4 4.Qh1 Rxd4
5.Qxf3 Qxb2 (8.373) 24

7.00 0:00 +1.35 1...Nf6 2.Qh3 Qd2 3.Rb1 Qxd4 4.Kh1 Rd5 (18.217) 48

8.00 0:00 +0.80 1...Nf6 2.Qh3 Qd2 3.Rb1 b6 4.Kh1 Rf4
5.Rbd1 Qxb2 (33.421) 82

9.00 0:00 +0.74 1...Nf6 2.Qh3 Qd2 3.Rb1 b6 4.Rfd1 Qc2
5.Rdc1 Qd2 (59.595) 127

10.00 0:00 +0.49 1...Nf6 2.Qh3 Qd2 3.Rb1 b6 4.Rfd1 Qc2
5.Rdc1 Qd3 6.d5 (119.140) 200

11.00 0:00 +0.52 1...Nf6 2.Qh3 Qd2 3.Rb1 Qc2 4.Qh2 b6
5.Rfc1 Qd2 6.Rd1 Qc2 (242.365) 287

11.19 0:03 +1.60 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qb6 6.Rf2 Qxd4 7.Bxb7 (1.315.015) 431

12.01 0:03 +1.78 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Rb8 7.Bxa6 Bc6 (1.751.409) 444

13.01 0:05 +1.94 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Rb8 7.Bg2 Qxd4+
8.Kh1 Qd2 (2.554.106) 464

14.01 0:08 +2.01 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Rb8 7.Bg2 Qxd4+
8.Kh2 Qg4 9.b4 (4.250.036) 485

15.01 0:17 +2.50 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Rb8 7.Bg2 g5 8.Rb1 Qxd4+
9.Kh1 gxh4 (8.983.659) 505

16.01 0:28 +2.35 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Rb8 7.Bg2 g5 8.Rb1 gxh4
9.Rf2 Qxd4 (14.626.149) 518

17.01 1:06 +2.17 1...Nxg3 2.Qxg3 Rf4 3.Bd5 Rg4 4.Bxf3 Rxg3+
5.fxg3 Qd2 6.Bxb7 Rb8 7.Bg2 g5 8.Rb1 gxh4
9.Rf2 Qxd4 10.Rd1 (34.886.721) 523


17.03 2:56 +3.64 1...Nf4 2.Kh1 Rh5 3.Qxf3 Rf8 4.Bd5 Rxh4+
5.gxh4 Qxh4+ 6.Kg1 Nh3+ 7.Kg2 Rxf3
8.Bxf3 Nf4+ 9.Kg1 Qg5+ 10.Kh1 b6
11.Ne4 Qh6+ 12.Kg1 Nh3+ 13.Kg2 Qf4 (92.662.991) 526

18.01 4:54 +4.07 1...Nf4 2.Kh1 Rh5 3.Qxf3 Rxh4+
4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Bd5 Rg6+
7.Qg3 Rxg3+ 8.fxg3 Qxg3+ 9.Kh1 Qh4+
10.Kg1 Bh3 11.Rxf4 Qxf4 12.Rd1 Bg4
13.Rf1 Qxd4+ 14.Kg2 (155.426.534) 528

19.01 7:33 +4.07 1...Nf4 2.Kh1 Rh5 3.Qxf3 Rxh4+
4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Rfd1 Nh3+
7.Kf1 Rf6 8.Rd3 Nxf2 9.Re3 Ng4
10.Ke2 Rxf3 11.Rxf3 Qh2+ 12.Kd3 Nf6 (242.309.592) 533

20.01 19:20 +4.29 1...Nf4 2.Kh1 Rh5 3.Qxf3 Rxh4+
4.gxh4 Qxh4+ 5.Kg1 Rc6 6.Bd5 Rg6+
7.Qg3 Rxg3+ 8.fxg3 Qxg3+ 9.Kh1 Qh4+
10.Kg1 Bh3 11.Rxf4 Qxf4 12.Rc2 Qxd4+
13.Rf2 b5 14.Be4 Qe3 (602.404.864) 518

best move: Nh5-f4 time: 25:45.031 min n/s: 516.616 nodes: 798.180.000


Not sure but I did not change my weights in the King Safety calculation, while in the default a factor two has gone missing. So it may be I am now using twice as much weight as before, depending on the contents in the SafetyTable[]. I have not checked those yet.

Regards, Eelco
Tord Romstad
Posts: 1808
Joined: Wed Mar 08, 2006 9:19 pm
Location: Oslo, Norway

Re: 080420? Hopelessly out of date: Here is 080422!

Post by Tord Romstad »

Hi Eelco,

Although I occasionally tell people about my test results in e-mails and private messages, I never report my own test results for Glaurung on public message boards. Tests reported on public forums should always be performed by independent testers, IMHO. Therefore, I ask you (and others!) to please not publish the results I tell you privately. Publishing the results of your own tests is fine, of course.

I'm of course not angry for you for posting my results in this particular case, as you couldn't know my view about this. Just please don't do it again in the future. :)

Tord
JVMerlino
Posts: 1407
Joined: Wed Mar 08, 2006 10:15 pm
Location: San Francisco, California

Re: Test position

Post by JVMerlino »

My new copy of CM GM prefers Nf4!?, which no other engine reported seems to have considered, after just over one minute on a P4-3.0.

Is it better?

Unfortunately, it appears that this new Chessmaster does not give the complete engine output that previous editions gave. Meaning, it does not give the PV for each successive search depth, and only reports a new PV when the best move changes. VERY annoying. :x Does anybody know a way around this?

Code: Select all

Time	Depth	Score	Positions	Moves
0:35	2/12	-0.66	5013599		1...Nf6 2.Qh3 Qd2 3.Rb1 Rb5 4.Be6
					Rxb2 5.Rxb2 Qxb2 6.Nd5 Nxd5 7.Bxd7
					Rf8 8.Bc8
5:03	2/13	-1.03	45149955	1...Nf4 2.Kh1 Rcf8 3.Qxf3 Bc6 4.Ne4
					Nh3 5.Rxc6 Qxc6 6.d5 Nxf2+ 7.Nxf2
					Qf6 8.Qe3 Rxf2 9.Rxf2 Qxf2 10.Qxf2
					Rxf2
jm
User avatar
Eelco de Groot
Posts: 4702
Joined: Sun Mar 12, 2006 2:40 am
Full name:   Eelco de Groot

Re: 080420? Hopelessly out of date: Here is 080422!

Post by Eelco de Groot »

Understood Tord. Apologies. The testers should have the final say, it is good to have a standardized way of testing, you can play 100 tournaments but if they are all played under different conditions it is difficult to compare them.

I hope you can make some progress with Glaurung! It is incredible how much code you have added to it in just two days, is there maybe a competition for that then I am sure that you will win it!

Curious how the new version will perform against Strelka and the other members of that family, well they are clones of Rybka I know but I am waiting for Rybka 3 to buy, maybe Naum is also a good program to test against. You know what I just skipped a night being so occupied with trying to figure out out why my program didn't work... sometimes a little, occasional, obsession is not so bad, maybe it should be periodic thing, a "binge" is that the right word? Had a lot of fun though!

Regards, Eelco