Evaluation challenge

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

Moderators: hgm, Rebel, chrisw

User avatar
Ovyron
Posts: 4556
Joined: Tue Jul 03, 2007 4:30 am

Re: Evaluation challenge

Post by Ovyron »

Ugh, I ran a bunch of engines on the position, several found it at Depth 7, some at Depth 6, and the rest were not noteworthy. Then Talkchess failed to load and I only got "confirm form resubmission" once, and Talkchess failed to load, and the post was lost in the ether.

But I guess this new A/B champion was all that mattered:

Komodo 10.1:

1.00 0:00 -2.27 1.a3 (20.018) 2223
1.00 0:00 -1.42 1.e4 (66.776) 4769
1.00 0:00 -0.22 1.Be4 (67.818) 4843
2.00 0:00 +0.19 1.Be4 Nxc1 2.Rxc1 (68.404) 4885
3.00 0:00 -0.14 1.Be4 Nf2 2.Qxf2 (69.111) 4606
4.00 0:00 +0.46 1.Be4 Nf2 2.Bxh7+ Kxh7 3.Qxf2 (92.367) 4197
5.00 0:00 +0.38-- 1.Be4 Nf2 (93.202) 4051
5.00 0:00 -0.11-- 1.Be4 Rxe4 (93.970) 4085
5.00 0:00 +0.08-- 1.b3 (94.133) 4092
5.00 0:00 -1.21-- 1.b3 Nxc1 (94.329) 4100
5.00 0:00 -0.95-- 1.a3 (115.553) 3984
5.00 0:00 -0.19-- 1.Rxd3 (158.039) 4270
5.00 0:00 -0.19 1.Rxd3 cxd3 2.Be4 Bc5 3.Rg1 (186.216) 4533

Engines that won't do it faster: Andscacs 0.93, CFish, Crystal, Equinox 3.20, Fizbo 2.0, Houdini 6, Ginkgo (Fritz 17), Gull 3, Komodo 11.3.1, McCain X, Naum 4.6, OpenTal, Rybka 5 (Fritz 15), Shredder 13, Shredder 12, Stockfish Polyglot, Strelka 5.5, SugaR NN, Zappa Mexico II.
User avatar
Rebel
Posts: 6991
Joined: Thu Aug 18, 2011 12:04 pm

Re: Evaluation challenge

Post by Rebel »

Current standing

Code: Select all

   Engine            Depth      Nodes
1. Lc0 [J13B.3-200]    1          6
1. Lc0 [11248]         1          6
3. Komodo 10           5
4. Lc0 [MeanGirl7]     6
5. Rodent IV           7
5. Amoeba 3.0          7
7. Critter 1.4         9
8. Stockfish-dev      10
9. Weakfish XR7       13
10 Minic              19
Best AB engine Komodo 10.

10 engines, poll closed.
90% of coding is debugging, the other 10% is writing bugs.
User avatar
Eelco de Groot
Posts: 4561
Joined: Sun Mar 12, 2006 2:40 am
Full name:   

Re: Evaluation challenge

Post by Eelco de Groot »

Ovyron wrote: Tue Dec 17, 2019 3:38 am Ugh, I ran a bunch of engines on the position, several found it at Depth 7, some at Depth 6, and the rest were not noteworthy. Then Talkchess failed to load and I only got "confirm form resubmission" once, and Talkchess failed to load, and the post was lost in the ether.

But I guess this new A/B champion was all that mattered:

Komodo 10.1:

1.00 0:00 -2.27 1.a3 (20.018) 2223
1.00 0:00 -1.42 1.e4 (66.776) 4769
1.00 0:00 -0.22 1.Be4 (67.818) 4843
2.00 0:00 +0.19 1.Be4 Nxc1 2.Rxc1 (68.404) 4885
3.00 0:00 -0.14 1.Be4 Nf2 2.Qxf2 (69.111) 4606
4.00 0:00 +0.46 1.Be4 Nf2 2.Bxh7+ Kxh7 3.Qxf2 (92.367) 4197
5.00 0:00 +0.38-- 1.Be4 Nf2 (93.202) 4051
5.00 0:00 -0.11-- 1.Be4 Rxe4 (93.970) 4085
5.00 0:00 +0.08-- 1.b3 (94.133) 4092
5.00 0:00 -1.21-- 1.b3 Nxc1 (94.329) 4100
5.00 0:00 -0.95-- 1.a3 (115.553) 3984
5.00 0:00 -0.19-- 1.Rxd3 (158.039) 4270
5.00 0:00 -0.19 1.Rxd3 cxd3 2.Be4 Bc5 3.Rg1 (186.216) 4533

Engines that won't do it faster: Andscacs 0.93, CFish, Crystal, Equinox 3.20, Fizbo 2.0, Houdini 6, Ginkgo (Fritz 17), Gull 3, Komodo 11.3.1, McCain X, Naum 4.6, OpenTal, Rybka 5 (Fritz 15), Shredder 13, Shredder 12, Stockfish Polyglot, Strelka 5.5, SugaR NN, Zappa Mexico II.
For older engines, maybe for most engines around that match it was a very hard problem I think. Fruit can solve it fast and possibly engines like Shredder before that but I have not tested any of the Shredder engines although I have several on this older machine from 2005. Some more oldies, Chess Tiger 2007.1 gambit could not find it at all, although in the end the eval dropped:


Chess Tiger 2007.1 Gambit 96 Mb

00:00:00.0 -1,6 2 1 11 Qd2
00:00:00.0 -1,74 2 546 Qd2 Rbd8 Rhg1
00:00:00.0 -1,14 2 784 Be4 Nxc1 Rxc1
00:00:00.0 -1,24 3 999 Be4 Nc5 Bc2
00:00:00.0 -1,38 4 2702 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxh1 Rxh1
00:00:00.0 -1,38 5 3714 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxh1 Rxh1
00:00:00.0 -1,6 6 8077 Be4 Nc5 Bc2 c3 e4 b4 h6
00:00:00.0 -1,38 7 21099 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxh1 Rxh1
00:00:00.0 -1,22 8 35964 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxh1 Rxh1 Rd8 Rg1
00:00:00.0 -1,22 9 53961 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxd1 Rxd1 Rd8 e4 Bxf4 Bxf4
Qxf4
00:00:00.1 -1,34 10 90517 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxd1 Rxd1 Re8 e4 Bxf4 Bxf4
Qxf4
00:00:00.2 -1,24 11 202045 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxd1 Rxd1 Re8 e4 Bxf4 Bxf4
Qxf4 Rd4
00:00:00.7 -0,80 12 577914 Be4 Nxc1 Rxc1 Qe7 Bf5 Qxe3 Rcg1 Qd4 h6 g6 Qc2 Qf6
Rg4
00:00:01.7 -0,92 13 1345083 Be4 Nxc1 Rxc1 h6 Qc2 Re7 Rcg1 Rbe8 Bh7+ Kh8 e4 Bc5
Bf5 Bxg1 Rxg1
00:00:02.7 -0,98 14 2110753 Be4 Nxc1 Rxc1 h6 Qc2 Re7 Rcd1 Rbe8 Bh7+ Kf8 e4 Bxf4
Rhg1 Be5 Bf5
00:00:06.5 -0,81 15 5212505 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxd1 Rxd1 Re8 Bd2 Qh4 Qf5 Qe7
h6 Qe4+ Qxe4 Rxe4 hxg7 Kxg7
00:00:08.9 -0,92 16 7232077 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxh1 Rxh1 b4 Qxc4 b3 axb3 Rb4
Qc8+ Bf8 Qc2
00:00:20.3 -0,82 17 16988581 Be4 Nxc1 Rxc1 Qe7 Bc2 b4 h6 g6 b3 c3 f5 Qf6
e4 Be5 fxg6 fxg6 Rhg1
00:00:47.1 -0,70 18 37750369 Be4 Nxc1 Rxc1 h6 Rh3 b4 Rxc4 b3 axb3 Rxb3
Qc2 Rb5 Ra4 Be5 fxe5 Qf1+ Ka2 Qxh3
00:01:19.5 -0,92 19 59818772 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxd1 Rxd1 Qh4 Rg1
Qxh5 Bd2 b4 Qxc4 Qh2 Qc1 Re8 Rg5 Bc5 Qd1
00:02:39.5 -0,70 20 124359638 Be4 Nxc1 Rxc1 Qe7 Bc2 b4 Rhg1 Qf6 h6 g6 Ba4
Rec8 Bc6 c3 Qg5 Qxg5 fxg5 Bc5
00:04:41.5 -0,62 21 211738343 Be4 Nxc1 Rxc1 Qe7 Bc2 b4 Rhg1 Qf6 h6 g6 Ba4
Rec8 Bc6 c3 Qg5 Qxg5 fxg5 Bc5
00:08:54.5 -0,62 22 399151983 Be4 Nxc1 Rxc1 Qe7 Bc2 b4 Rhg1 Qf6 h6 g6 Ba4
Rec8 Bc6 c3 Qg5 Qxg5 fxg5 Bc5
00:17:34.8 -0,84 23 840180805 Be4 Nxc1 Rxc1 Qe7 Bc2 b4 Rhg1 Qf6 h6 g6 Ba4
Rxe3 Rxc4 Bxf4 Qf2 Bg5 Qxf6 Bxf6 Rgc1 Rd8 Rxb4 Rxd5 Rc8+
00:54:36.7 -0,64 24 24 94365915 Be4 Nxc1 Rxc1 Qe7 Bc2 b4 Rh3 b3 axb3 cxb3
Bd3 Rb4 Rg1 Qf6 Rg3 Bf8 e4 Qd4
03:17:43.9 -0,86 25 2741714981 Be4 Nxc1 Rxc1 Rbc8 h6 g6 f5 Qe5 Rh4 c3 Rg4
cxb2 Rc6 Rxc6 dxc6 Qc3 Qxb2 Qe1+ Qc1

I apologize for the lay out, there are no spaces between eval and iteration number after transferring the Chess Partner output to Notepad to here and when I add some spaces they almost do disappear again in the forum layout.

Rybka 1.0 beta was fast:

Code: Select all

Rybka 1.0 Beta (Very Positional, 64 Mb)

00:00:00.1	-0,77 	3	207	Be4 
00:00:00.1	-0,73	        3	356	h6 
00:00:00.1	-0,68	        3	414	Rxd3 
00:00:00.1	-0,71  	4	747	Rxd3 
00:00:00.1	-0,61	        4	2355	Qc2 
00:00:00.2	-0,37         5	7191	Qc2 Nxc1 
00:00:00.3	-0,48   	6	11897	Qc2 Nxc1 Qxc1 b4 
00:00:00.5	-0,69         7	23538	Qc2 Nxc1 Qxc1 c3 bxc3 b4 
00:00:00.6	-0,52	         7	29803	Rxd3 cxd3 Be4 Bc5 
00:00:00.9	-0,22	         8	52822	Rxd3 cxd3 Be4 Bc5 Rh3 
00:00:01.3	-0,27         9	82113	Rxd3 cxd3 Be4 Bc5 Rh3 b4 
00:00:02.3	-0,14	       10	149911	Rxd3 cxd3 Be4 Bc5 Rh3 Rbc8 Bxd3 
00:00:04.6	-0,02       11	330717	Rxd3 cxd3 Be4 b4 Bxd3 b3 axb3 h6 Bc2 
00:00:08.2	-0,08	       12	626404	Rxd3 cxd3 Be4 Bc5 Rh3 Rbc8 Bxd3 b4 Bd2 
00:00:15.5	-0,10	       13	1239368	Rxd3 cxd3 Be4 Rbc8 Bxd3 h6 Rg1 b4 Bd2 a5 
00:00:28.5	-0,07	       14	2337573	Rxd3 cxd3 Be4 Rbc8 Bxd3 h6 Rg1 b4 Ka1 Kf8 
00:01:00.2	-0,02	       15	4962227	Rxd3 cxd3 Be4 Rbc8 Bxd3 h6 Rg1 a6 Qg4 Bb4 
00:02:39.8	0,01	        16	12651378	Rxd3 cxd3 Be4 b4 Bxd3 Qh6 b3 Rbc8 Bd2 Rc7 
00:04:59.2	0,00	        17	23611187	Rxd3 cxd3 Be4 b4 Bxd3 Rbc8 h6 g6 Bd2 a5 
00:09:37.7	0,00	        18	47060545	Rxd3 cxd3 
Fruit Beta from 2005, from the time Fabien thought about making fruit commercial, this was about as strong as rybka 1.0 beta but it was not tested enough so they only found out later it was strong

Code: Select all

Fruit Beta 05/11/03 (256 Mb. This is with the default settings) 05 is 2005 I think, that is the date I have for the exe, november 2005

00:00:00.0	-0,64	       1	319	Be4 
00:00:00.0	-0,35       2	524	Be4 Rbc8 
00:00:00.0	-0,38	       3	1103	Be4 Rbc8 Ka1 
00:00:00.0	0,01	       4	2365	Be4 Rbc8 h6 g6 Bxd3 cxd3 Rxd3 
00:00:00.0	0,01	       5	4306	Be4 Rbc8 h6 g6 Bxd3 cxd3 Rxd3 
00:00:00.1	-0,83	       6	15084	Be4 Nxc1 Rxc1 Qe7 Bf5 Qxe3 h6 
00:00:00.2	-0,40	       6	48171	Qc2 Nxc1 Qxc1 Bc5 e4 a5 h6 
00:00:00.2	-0,33	       7	98620	Qc2 Nb4 Qg2 Bxf4 a3 Bxe3 axb4 
00:00:00.4	-0,38	       8	212849	Qc2 Nxc1 Qxc1 Rbc8 Ka1 c3 Rd3 cxb2+ Qxb2 Qxb2+ Kxb2 
Rc4 Rg1 
00:00:00.8	-0,39	       9	429741	Qc2 Nb4 Qg2 Bc5 Bd2 Rxe3 Bxe3 Bxe3 h6 g6 Qg5 Qxg5 
fxg5 Bxg5 
00:00:02.0	-1,59	      10	937612	Qc2 Bxf4 Rhf1 Nxc1 Be4 Nd3 Bxd3 Rxe3 Bxh7+ Kf8 Qg2 
00:00:02.3	-0,60       10	1140471	Be4 Nxc1 Rxc1 Qe7 Bf5 Qxe3 Rcg1 Bf8 Qg5 Rb6 Ka1 
00:00:02.9	-0,36	      10	1481616	Rxd3 cxd3 Be4 Bc5 Rh3 b4 Bxd3 b3 axb3 a5 Qg5 Qxg5 
fxg5 Rxb3 
00:00:03.7	-0,28	      11	1945820	Rxd3 cxd3 Be4 b4 Bxd3 b3 axb3 a5 Bc2 a4 Bd2 Rec8 
00:00:05.5	-0,12      12	3060950	Rxd3 cxd3 Be4 b4 Bxd3 b3 axb3 a5 Rg1 Rec8 Bd2 Rxb3 
Bc3 
00:00:09.1	-0,08	       13	5227274	Rxd3 cxd3 Be4 b4 Bxd3 b3 axb3 a5 Rg1 Rec8 Qh3 Rc7 
h6 g6 
00:00:17.4	-0,10	       14	10160648	Rxd3 cxd3 Be4 Bc5 Rh3 b4 Bxd3 b3 axb3 Rec8 
h6 g6 Bc4 a5 Bd2 
00:00:34.0	-0,04	       15	20572426	Rxd3 cxd3 Be4 Bc5 Rh3 Bb4 Bxd3 a6 a3 Be7 h6 
g6 Bd2 Rbc8 Bc3 Rxc3 bxc3 Qxc3 
00:01:20.8	-0,07	       16	47906014	Rxd3 cxd3 Be4 Qh6 Bxd3 b4 Bc2 Rbc8 Bd3 b3 
axb3 Rb8 Bc2 Bc5 Qg3 Rbd8 b4 Bxb4 
00:02:24.6	0,00	      17	87858323	Rxd3 cxd3 Be4 Qh6 Bxd3 b4 Bc2 Rbc8 Bd3 Rb8 
00:06:51.9	0,08	      18	255329715	Rxd3 cxd3 Be4 Qh6 Bxd3 b4 Bd2 b3 a3 Rbc8 
Rg1 Qf6 Qg5 Qxg5 Rxg5 Rc5 h6 g6 Bb4 Rxe3 Bxc5 Bxc5 
00:12:44.7	0,08	       19	476592883	Rxd3 cxd3 Be4 Qh6 Bxd3 b4 Bd2 b3 a3 Rbc8 
Qg5 Qxg5 fxg5 Rc7 h6 g6 e4 Be5 Be3 Rec8 Rf1 
Rebel 12 also does not find it:

Rebel 12

00:00:00.2 -10,12 1 26 Qg6 fxg6
00:00:00.2 -1,76 1 38 Rxd3 cxd3
00:00:00.2 -1,12 1 76 e4 Bxf4
00:00:00.2 -0,30 1 78 Qg5
00:00:00.2 -0,01 1 182 Be4
00:00:00.2 -0,14 2 460 Be4 Rbc8
00:00:00.2 -0,10 3 1461 Be4 Rbc8 Rhg1
00:00:00.3 -0,09 4 8455 Be4 Rxe4 Qxe4 Nf2 Qc2 Nxh1
00:00:00.3 -0,38 5 21043 Be4 Nxc1 Kxc1 c3 Rd4 cxb2+ Qxb2 Rbc8+ Kb1
00:00:00.4 -0,36 6 98086 Be4 Nxc1 Kxc1 c3 Rd4 cxb2+ Qxb2 Rbc8+ Kb1 a6
00:00:00.5 -0,60 7 223487 Be4 Nxc1 Kxc1 c3 Rd4 cxb2+ Kb1 Bc5 Rd3
00:00:01.2 -0,58 7 882298 Qc2 Bxf4 exf4 Nb4 Qg2 Qf5+ Be4 Rxe4
00:00:01.6 -1,44 8 1249781 Qc2 Bxf4 Rhf1 Qa6 Rxd3 cxd3 Qxd3 Be5 Bg4 Qa4
00:00:01.7 -0,53 8 1386159 Be4 Nxc1 Kxc1 c3 b4 Rbc8 Rd4 Kf8 Rhd1
00:00:02.8 -0,57 9 2351676 Be4 Nxc1 Kxc1 c3 Kb1 cxb2 h6 g6 Rd2 Qe7 Rd4
00:00:05.0 -0,57 10 4681226 Be4 Nxc1 Kxc1 c3 Kb1 cxb2 h6 g6 Rd2 Qe7 Rd4
00:00:08.1 -0,69 11 7784056 Be4 Nxc1 Rxc1 Qe7 Bf5 Qxe3 Rhg1 Qd4 Rcd1 Qf6 Qc2
00:00:24.5 -0,67 12 25141504 Be4 Nxc1 Rxc1 Qe7 Bf5 Qxe3 Rhg1 Qd4 Qc2 h6
Bh7+ Kf8
00:00:31.6 -0,62 12 32729995 e4 Bxf4 Bxf4 Qxf4 h6 g6 Rdf1 Qe5 Bg4 b4
00:00:51.7 -0,65 13 54987594 e4 Bxf4 Bxf4 Nxf4 Qg3 Nd3 Rd2 Rbc8
00:00:57.2 -0,62 13 61444585 Be4 Nxc1 Rxc1 Qe7 Bc2 Qxe3 h6 g6 f5 g5 f6
Qf4
00:02:05.2 -0,59 14 136815407 Be4 Nxc1 Rxc1 Qe7 Bf5 Qxe3 Rhg1 Qd4 Rcd1
Qf6 Qc2
00:03:59.1 -0,55 15 267587738 Be4 Nxc1 Rxc1 Qe7 Bc2 Qxe3 Rcg1
00:13:21.3 -0,52 16 938783812 Be4 Nxc1 Rxc1
00:46:03.1 -0,50 17 3220391049 Be4 Nxc1
03:31:52.5 -0,48 18 1046517549 Be4 Nxc1


However, after some trying I did manage to find some settings for Rebel 12 that eventually does play Rxd3! Not quickly but just finding it was nice :D

00:00:00.1 -10,23 1 27 Qg6 hxg6
00:00:00.1 -1,76 1 39 Rxd3 cxd3
00:00:00.1 -1,14 1 77 e4 Bxf4
00:00:00.1 -0,30 1 79 Qg5
00:00:00.1 -0,05 1 183 Be4
00:00:00.1 -0,08 2 526 Be4 Rbc8 Bxd3 cxd3 Rxd3
00:00:00.2 -0,08 3 1627 Be4 Rbc8 Bxd3 cxd3
00:00:00.2 -0,04 4 11457 Be4 Rxe4 Qxe4 Nf2 Qf3 Nxh1
00:00:00.2 -0,28 5 23169 Be4 Nxc1 Kxc1 c3 Rd4 cxb2+ Qxb2 Rbc8+ Kb1
00:00:00.3 -0,26 6 71236 Be4 Nxc1 Kxc1 c3 Rd4 cxb2+ Qxb2 Rbc8+ Kb1 a6
00:00:00.5 -0,48 7 195979 Be4 Nxc1 Kxc1 c3 Rd4 cxb2+ Kb1 Bc5 Rd3
00:00:01.0 -0,35 7 641719 Qc2 Rbd8 h6 g6 Be2 Nb4
00:00:01.6 -1,33 8 1273175 Qc2 Bxf4 Rhf1 Qa6 Qg2 Nxc1 Rxc1 Bxe3 Rcd1
00:00:01.7 -0,66 8 1405619 Be4 Nxc1 Kxc1 c3 h6 cxb2+ Kb1 g6 Rd2 Ba3
00:00:04.7 -0,59 9 3854602 Be4 Nxc1 Rxc1 Qe7 Bf5 Qxe3 Rhg1 Bf8 Rcd1 Qxf4
00:00:07.4 -0,59 10 6681903 Be4 Nxc1 Rxc1 Qe7 Bf5 Qxe3 Rhg1 Bf8 Rcd1 Qxf4
00:00:10.4 -0,66 11 9457842 Be4 Nxc1 Kxc1 c3 Kb1 cxb2 Rd4 Bc5 Rd3 Rbd8 h6 g6
00:00:27.5 -0,69 12 26355326 Be4 Nxc1 Kxc1 c3 h6 cxb2+ Kb1 g6 Rd4 Bc5 Rd3 Rbd8 Rhd1
00:00:35.2 -0,63 12 33819192 e4 Bxf4 Bxf4 Qxf4 h6 g6 Be2 f5 Bxd3
00:00:57.1 -0,71 13 56046936 e4 Bxf4 Bxf4 Qxf4 Rhg1 Qg5 Qxg5 Nxb2
00:01:04.8 -0,65 13 64005253 Be4 Nxc1 Rxc1 Qe7 Bf5 Qxe3 Rhg1 Qd4 Rcf1 Qf6
00:02:02.3 -0,62 14 125105467 Be4 Nxc1 Rxc1 Qe7 Bc2 Qxe3 Rhg1 Qd4 Rcd1
00:05:27.6 -0,69 15 342872271 Be4 Nxc1 Rxc1 Qe7 Bc2
00:07:01.6 -0,61 15 435903875 e4 Bxf4 Bxf4 Qxf4 Rdf1 Qe5 h6 g6 Rh2 b4 Bh5 Qxe4 Be2
00:17:22.7 -0,67 16 1053750251 e4 Bxf4 Bxf4 Qxf4 Rh3 Rb7 Rg3 f5
00:27:42.7 -0,57 16 1754699057 Rxd3 cxd3 Be4 Rbc8

Code: Select all

[Personality = BETAZOID II.ENG]       * based on Rebel 12 default engine
[Pawn Value = 100]    
[Knight Value = 100]  
[Bishop Value = 98] 
[Rook Value = 102]   
[Queen Value = 105]             * 105
[King Safety = 110]             * 139
[Mobility = 120]                * 120
[Pawn Structure = 100]          * 95
[Passed Pawns = 140]            * 105
[Passed Pawn King Tropism = 120]  
[Traditional Isolated Pawns = 80] 
[Progressive Isolated Pawns = 125]
[Pins = 110]                    * 110
[Bishop Pair = 140]       
[Chess Knowledge = 300]         * 200
[Attractiveness = 120]          * 119
[Attacking = 110]               * 125
[Strength of Play = 100]
[Draw Contempt Factor = 0.00]
[Selective Search = 200]
[Search Technique = NULLMOVE]

[Engine Learner = off]          * off|on
[Book Learner = aggressive]     * off|passive|moderate|strong|aggressive
[Position Learner = on]         * off|on
[Extended Book Learner = off]   * off|read|write|read&write

[Pruning = MISC_03]           
[Pruning = MISC_29]        
[Pruning = MISC_49]        

[Pruning = MISC_09]        
[Pruning = MISC_10]        
[Pruning = MISC09_3.00]   
[Pruning = MISC10_5.00]   

[Pruning = MISC_17]       
[Pruning = MISC17_DEPTH_10]

[Search Safety MIDG = 200]     
[Search Safety END0 = 250]    
[Search Safety END1 = 300]      
[Search Safety END2 = 300]      

[Center Control = 25]    
[Bishop Mobility = 25]   
[Right to Move = 75]     
[Strong Squares = 125]   

[ANTI-GM = OFF]
[EVALUATION = NORMAL]  
[Extensions (remaining)= 3]
[Extensions (checks)= 0]
[Extensions (captures)= 0]

Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
User avatar
Eelco de Groot
Posts: 4561
Joined: Sun Mar 12, 2006 2:40 am
Full name:   

Re: Evaluation challenge

Post by Eelco de Groot »

Betazoid II is holding on to Rxd3! Depth 17, after 1 hour 42 minutes

01:42:53.8 -0,41 17 1864586045 Rxd3 cxd3 Rf1 Rbd8
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
User avatar
Eelco de Groot
Posts: 4561
Joined: Sun Mar 12, 2006 2:40 am
Full name:   

Re: Evaluation challenge

Post by Eelco de Groot »

Eelco de Groot wrote: Tue Dec 17, 2019 11:01 pm Betazoid II is holding on to Rxd3! Depth 17, after 1 hour 42 minutes

01:42:53.8 -0,41 17 1864586045 Rxd3 cxd3 Rf1 Rbd8
Unfortunately, this was also not yet completely stable because Betazoid II went back to Be4, but it lasted well into the 18th iteration.

.
.
00:07:01.6 -0,61 15 435903875 e4 Bxf4 Bxf4 Qxf4 Rdf1 Qe5 h6 g6 Rh2 b4 Bh5 Qxe4 Be2
00:17:22.7 -0,67 16 1053750251 e4 Bxf4 Bxf4 Qxf4 Rh3 Rb7 Rg3 f5
00:27:42.7 -0,57 16 1754699057 Rxd3 cxd3 Be4 Rbc8
01:42:53.8 -0,41 17 1864586045 Rxd3 cxd3 Rf1 Rbd8
06:45:16.6 -0,41 18 380953559 Rxd3 cxd3
08:20:30.3 -0,39 18 2621681550 Be4 Nxc1

For a better eval of the Be4 variation, this is Hiarcs X54 Hypermodern II, with a little bit of learning because position learning was on, after 1. Be4 Nxc1 2. Rxc1 Qe7 3. Bc2 Qxe3 4. Rhg1 Bf8

[pgn][Event "?"] [Site "?"] [Date "2019.12.9"] [Round "?"] [White "Eelco de Groot"] [Black "Chess Tiger 2007.1 engine - Gambit style"] [Result "*"] [Setup "1"] [FEN "1r2r1k1/p4ppp/3b1q2/1p1P3P/2p2P2/3nPB2/PP4Q1/1KBR3R w - - 0 1"] 1. Rxd3 ( 1. Be4 Nxc1 2. Rxc1 Qe7 3. Bc2 Qxe3 4. Rhg1 Bf8 ) 1... cxd3 *[/pgn]


[d]1r2rbk1/p4ppp/8/1p1P3P/2p2P2/4q3/PPB3Q1/1KR3R1 w - - 2 1

Three best moves, eval for White:

02:31:33.4 -1,08 18 2240008727 d6 Rbd8 Qg5 Rxd6 Qxb5 c3 Qf5 Rh6 Bb3 Re7 bxc3 Rc6 Rge1 Qxe1 Rxe1 Rxe1+ Kb2 Rf6 Qa5 Rxf4 Qxa7 Re2+
02:31:34.5 -1,08 18 2240008727 Qg5 Rb6 d6 Rxd6 Qxb5 c3 Qf5 Rh6 Bb3 Re7 bxc3 Rc6 Rge1 Qxe1 Rxe1 Rxe1+
02:31:34.5 -1,13 18 2240008727 Rcd1 b4 Qg5 b3 axb3 cxb3 Bd3 Rb6 Qf5 g6 hxg6 fxg6 Qd7 Rd6 Qc7 Qf2 Rgf1 Qb6 Qxb6 Rxb6 Rde1 Rd8

better spacing:

Code: Select all

02:31:33.4	-1,08 	          18	2240008727	d6 Rbd8 Qg5 Rxd6 Qxb5 c3 Qf5 Rh6 Bb3 Re7 bxc3 Rc6 Rge1 Qxe1 Rxe1 Rxe1+ Kb2 Rf6 Qa5 Rxf4 Qxa7 Re2+ 
02:31:34.5	-1,08	          18	2240008727	Qg5 Rb6 d6 Rxd6 Qxb5 c3 Qf5 Rh6 Bb3 Re7 bxc3 Rc6 Rge1 Qxe1 Rxe1 Rxe1+ 
02:31:34.5	-1,13	          18	2240008727	Rcd1 b4 Qg5 b3 axb3 cxb3 Bd3 Rb6 Qf5 g6 hxg6 fxg6 Qd7 Rd6 Qc7 Qf2 Rgf1 Qb6 Qxb6 Rxb6 Rde1 Rd8
Debugging is twice as hard as writing the code in the first
place. Therefore, if you write the code as cleverly as possible, you
are, by definition, not smart enough to debug it.
-- Brian W. Kernighan
User avatar
lucasart
Posts: 3232
Joined: Mon May 31, 2010 1:29 pm
Full name: lucasart

Re: Evaluation challenge

Post by lucasart »

Rebel wrote: Mon Dec 16, 2019 10:20 am [d]1r2r1k1/p4ppp/3b1q2/1p1P3P/2p2P2/3nPB2/PP4Q1/1KBR3R w - -
This is a slightly modified position from the game Anand - Rebel 10 (1998).

Rebel just had played 1..Nd3 with a happy score.

Much to my surprise Anand answered in a reflex and removed the strong knight from the board with 2.Rxd3 within a second. A moment of learning something new about chess.

The challenge then: The best engine (out of the box, no special settings) will be that plays 2.Rxd3 at the earliest depth possible and remains stable for 1 minute.
Demolito found it almost immediately:

Code: Select all

run 1:
info depth 17 score cp -75 time 2059 nodes 22312353 hashfull 27 pv d1d3

run 2:
info depth 15 score cp -55 time 356 nodes 3545315 hashfull 4 pv d1d3
Doing 2 runs, because SMP search is not deterministic (second run luckier).

But what makes you so certain that this is the only correct move ? It is a fine move, and is easy for humans to understand, but possibly not the only one that holds the draw (with perfect play).

PS: Moron 1.0 plays Bd2 :lol:
Theory and practice sometimes clash. And when that happens, theory loses. Every single time.
PK
Posts: 893
Joined: Mon Jan 15, 2007 11:23 am
Location: Warsza

Re: Evaluation challenge

Post by PK »

Moron 0.9 played 1.Qf2. Glad it has been fixed.