Shirov's Bh3, Houdini solves it

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

Moderator: Ras

peter
Posts: 3477
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: Shirov's Bh3, Houdini solves it

Post by peter »

Hi Uri!
Uri Blass wrote:
[d]8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47

Stockfish_16121212_x64_modern:
63/73 05:17 1,363,706k 4,296k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bd4 Kh4 77.Bb2 Bc2 78.Kg2 Kg4 79.Be5 Bb1 80.Kf1 Kh3 81.Kg1 Be4 82.Bb2 Kg4 83.Kf1 Kf4
In this Output- line latest 51.Kf2?? is a crucial blunder, after that game's lost, with 51.Bxf6 it is not,
Peter.
Top
Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 12:41 pm

Re: Shirov's Bh3, Houdini solves it

Post by Lyudmil Tsvetkov »

Uri Blass wrote:
Lyudmil Tsvetkov wrote:
peter wrote:Dead draw, Lyudmil.

BTW, don't you think that to be a fittle funny too:
...do not look at what engines suggest...
...I ran a 10 games shootout with Komodo 10.1...
...games were at 1 min. bullet...
:)
I do not see what is so funny; you migth play 100 instead of 10 games, and black will still score around 30% wins or so. That is how engines are tuned, using stats of winning probabilities.

objectively migth be drawn (which I am still not certain), but black has nice winning chances after Kf5, capturing the h pawn and going to the first rank on f1.

problem however is not if this is a draw or win, but, as John wanted to know, why top engines are unable to find the key move. I myself do not see how an accurate eval assessment of this position could be returned without specifying some 500 lines of code, because:

- most eg positions with 2 passers and 3 files apart migth be drawn, but with c & f files wins
- positions with pawn more than 3 files apart might be wins, but with edge a or h pawns and wrong bishop draws; still, even some of those are wins
- many things will also depend on the placement of remaining pawns, some positions are draws, while others wins

so, if engines give +200cps for most of those, how do you make the distinction?
The problem is not that the engine has an evaluation that is not accurate.

It is ok not to know if the position is a win for black and to think 30% win for black and 70% draw but in this case I could expect an evaluation of something like +0.6 for black.

The main problem is that the engine claim that it is almost sure black is winning based on the evaluation when it suggest a drawing move.

I think that it is bad that people use tablebases because tablebases hide the bugs of chess engines so sometimes people even do not see how wrong they can be.

Latest version of stockfish from 12.12(without tablebases) show the following analysis and I wonder what cause it to fail high again and again on a3 and when it is going to be stopped(6.54 pawns at depth 64 after more than an hour but I do not believe that a3 is winning):

[d]8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47

Stockfish_16121212_x64_modern:
1/1 00:00 35 35k +1.79 47. ... a3
2/2 00:00 102 102k +2.17 47. ... a3 48.g3
3/3 00:00 177 177k +2.52 47. ... a3 48.Ba1 a2
4/4 00:00 262 262k +2.52 47. ... a3 48.Ba1 a2 49.g3
5/6 00:00 2k 788k +1.78 47. ... a3 48.Kf2 Be4 49.Bd4 a2
6/6 00:00 4k 1,343k +1.73 47. ... a3 48.g3 a2 49.Kf2 Bd3 50.Kf3 Be4+ 51.Ke2
7/7 00:00 5k 1,606k +1.70 47. ... a3 48.g3 a2 49.Kf2 Be4 50.Bd4 f5
8/8 00:00 8k 2,013k +1.64 47. ... a3 48.g3 Bc2 49.Kf2 a2 50.Bd4 f5 51.Ke3
9/11 00:00 25k 2,779k +1.48 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 a2 51.Bd4 Bb3 52.Kd3
10/12 00:00 34k 2,801k +1.45 47. ... a3 48.Kf2 Be4 49.g3 Bc2 50.Ke3 Bf5 51.Kd4 Bc2 52.Kc5 a2
11/15 00:00 64k 2,922k +1.47 47. ... a3 48.Kf2 Be4 49.g3 Kf5 50.Ba1 Bc2 51.Ke3 Bb3 52.Bd4 Bc4 53.Kd2 a2
12/16 00:00 101k 2,980k +1.42 47. ... a3 48.Kf2 Bc2 49.Ke2 Kf5 50.Kd2 Bb1 51.g3 Kg4 52.Bxf6 Kxg3 53.Kc1 Be4
13/16 00:00 120k 3,073k +1.41 47. ... a3 48.Kf2 Bd3 49.g3 Kf5 50.Ke3 Bb1 51.Ke2 Kg4 52.Bxf6 Kxg3 53.Kd2 Kf4 54.Kc1
14/16 00:00 122k 3,049k +1.41 47. ... a3 48.Kf2 Bc2 49.g3 Kf5 50.Ke3 Bb1 51.Ke2 Kg4 52.Bxf6 Kxg3 53.Kd2 Kf4 54.Kc1
15/23 00:00 228k 3,504k +1.41 47. ... a3 48.Kf2 Be4 49.g3 f5 50.Ke3 a2 51.Kd4 Kd6 52.Bb2 Bf3 53.Kc3 Be2 54.Kb3 Bc4+ 55.Kc2
16/23 00:00 250k 3,577k +1.43 47. ... a3 48.Kf2 Be4 49.g3 f5 50.Ke3 Kd7 51.Kd2 Kc6 52.Bf6 Kb5 53.Kc3 Ka4 54.Bg7 Bh1 55.Kc2 Bg2 56.Bd4 Kb4
17/24 00:00 619k 3,558k +1.68 47. ... Be4 48.g4 a3 49.h5 gxh5 50.gxh5 f5 51.Kf2 f4 52.Bd4 a2 53.Ke1 Kd6 54.Kd1 f3 55.Ke1 Kc6 56.h6 Kb5
18/24 00:00 669k 3,503k +1.69 47. ... Be4 48.g4 a3 49.Kf2 f5 50.gxf5+ Kxf5 51.Ke3 Kg4 52.Bf6 a2 53.Kd2 Bb1 54.Bc3 Kxh4 55.Kc1 g5 56.Bf6 Kh3 57.Be5 g4
19/29 00:00 919k 3,283k +1.80 47. ... Be4 48.g4 a3 49.Kf2 f5 50.gxf5+ Kxf5 51.Ke3 a2 52.Bd4 Kg4 53.Kd2 Kxh4 54.Kc1 g5 55.Kb2 g4 56.Kxa2 g3 57.Kb3 g2 58.Kb4 Kg4
20/31 00:00 1,356k 3,095k +2.04 47. ... Be4 48.g4 a3 49.h5 gxh5 50.gxh5 f5 51.Kf2 f4 52.Ke2 Kf5 53.h6 Kg5 54.Bg7 a2 55.Kf2 Kg4 56.Ke2 Bg6 57.Bd4 Kf5 58.Kf3 Kg5 59.h7 Be4+ 60.Ke2 Bxh7
21/33 00:00 2,072k 3,465k +2.14 47. ... Be4 48.g4 a3 49.h5 gxh5 50.gxh5 f5 51.Kf2 f4 52.h6 a2 53.Ke2 Bg6 54.Kf3 Kf5 55.Bd4 Kg5 56.Bg7 Be4+ 57.Ke2 Kh5 58.Kf2 Kg4 59.Bd4 Bg6 60.Bg7 f3 61.Ke3 Kg3 62.Be5+ Kg2
22/34 00:00 2,532k 3,617k +2.07 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Ba1 a2 56.Bf6 Bf5 57.Bd4 Bh3 58.Bf6 Bd7 59.Ba1
23/34 00:00 3,076k 3,715k +2.00 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Bf6 a2 56.Ba1 Bf5 57.Bd4 Kb3 58.Kd2 Bd7 59.Ba1 Kc4 60.Ke3 Bg4 61.Bf6
24/34 00:00 3,609k 3,815k +1.93 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 a2 56.Bf6 Bf5 57.Bd4 Bd7 58.Ba1 Bh3 59.Bf6 Bg2 60.Bd4 Bh3 61.Ba1 Bc8
25/34 00:01 3,862k 3,847k +1.93 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 a2 56.Bf6 Bf5 57.Bd4 Bd7 58.Ba1 Bh3 59.Bf6 Kb4 60.Kd2 Be6 61.Kc1 Kb3 62.Ba1
26/35 00:01 4,237k 3,894k +2.00 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bd4 Kb3 57.Kd2 Bf3 58.Ke3 Be4 59.Kd2 Kc4 60.Ke3 a2 61.Bf6 Bf5 62.Bd4 Bd7 63.Bf6
27/35 00:01 5,252k 3,817k +1.92 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 a2 56.Bf6 Bf5 57.Bd4 Bh3 58.Bf6 Kc5 59.Bc3 Be6 60.Bd4+ Kc4 61.Bf6 Kb4 62.Ba1
28/35 00:01 5,968k 3,765k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 Kb4 57.Bd4 Kb3 58.Kd2 Bf3 59.Ke3 Be4 60.Kd2 Kc4 61.Ke3 Kb4 62.Be5 a2 63.Kd2 Kc4 64.Kc1 d4
29/35 00:01 6,922k 3,793k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 a2 57.Bd4 Be4 58.Be5 Bb1 59.Bf6 Bf5 60.Be5 Bd7 61.Bf6 Be6 62.Be5 Bg4 63.Bf6 Bh3 64.Ba1 Kb3
30/35 00:02 7,987k 3,853k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 a2 57.Bd4 Be4 58.Be5 Bb1 59.Bf6 Bf5 60.Be5 Bd7 61.Bf6 Be6 62.Be5 Bg4 63.Bf6 Bh3 64.Ba1 Kb3
31/37 00:02 9,220k 3,802k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 a2 57.Bd4 Be4 58.Be5 Bb1 59.Bf6 Bf5 60.Be5 Bd7 61.Bf6 Be6 62.Be5 Bg4 63.Bf6 Kc5 64.Bd4+ Kb4 65.Ba1 Kc4
32/40 00:02 9,938k 3,793k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 Bh3 57.Bd4 Kb3 58.Kd2 Bg4 59.Bf6 Kc4 60.Ke3 Be6 61.Bd4 Bf5 62.Be5 Be4 63.Bd4 Bh1 64.Bf6 Bg2 65.Bd4 Kb3 66.Kd2
33/40 00:03 12,830k 3,837k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 Bh1 57.Bd4 Kb3 58.Kd2 Bf3 59.Bf6 Kc4 60.Ke3 Bg4 61.Ba1 Bd7 62.Bd4 Bh3 63.Be5 Be6 64.Bf6 Bf5 65.Be5 Bg4
34/40 00:03 15,124k 3,877k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 Bh1 57.Bd4 Kb3 58.Kd2 Bf3 59.Bf6 Kc4 60.Ke3 Bg4 61.Ba1 Bd7 62.Bd4 Bh3 63.Be5 Be6 64.Bf6 Bf5 65.Be5 Be4 66.Bf6
35/45 00:05 21,185k 3,895k +1.91 47. ... Be4 48.g4 f5 49.gxf5+ Bxf5 50.Kf2 Kd6 51.Bb4+ Kc6 52.Ke3 Kb5 53.Be7 Kc4 54.Kd2 Kb3 55.Bd6 a3 56.Be5 Kc4 57.Ke3 Bh3 58.Bd4 Be6 59.Ba1 Bd7 60.Bd4 Bh3 61.Be5 Kb3 62.Kd2 Bf5 63.Bf6 Kc4 64.Ke3 Bb1 65.Bd4 Be4 66.Be5 Bg2
36/47 00:06 25,725k 3,952k +1.91 47. ... Be4 48.g4 f5 49.gxf5+ Bxf5 50.Kf2 Kd6 51.Bb4+ Kc6 52.Ke3 Kb5 53.Be7 Kc4 54.Kd2 Kb3 55.Bd6 a3 56.Be5 Kc4 57.Ke3 Bh3 58.Bd4 Be6 59.Be5 Bd7 60.Bd4 Bh3 61.Be5 Kb3 62.Kd2 Bf5 63.Bf6 Kc4 64.Ke3 Bb1 65.Bd4 Be4 66.Be5 Bg2 67.Bf6 Bh1 68.Bd4 Kb3 69.Kd2
37/47 00:07 28,701k 4,002k +1.91 47. ... Be4 48.g4 f5 49.gxf5+ Bxf5 50.Kf2 Kd6 51.Bb4+ Kc6 52.Ke3 Kb5 53.Be7 Kc4 54.Kd2 Kb3 55.Bd6 a3 56.Be5 Kc4 57.Ke3 Bh3 58.Bd4 Be6 59.Be5 Bd7 60.Bd4 Bh3 61.Be5 Kb3 62.Kd2 Bf5 63.Bf6 Kc4 64.Ke3 Bb1 65.Bd4 Be4 66.Be5 Bg2 67.Bf6 Bh1 68.Bd4 Kb3 69.Kd2 Bf3
38/47+ 00:08 34,047k 4,048k +1.99 47. ... Be4
38/47 00:10 41,473k 4,127k +1.92 47. ... Be4 48.Kf2 f5 49.g3 Kd6 50.Bd4 Kc6 51.Ke2 a3 52.Kd2 a2 53.Kc1 Kb5 54.Kd2 Kc4 55.Ke3 Kb3 56.Kd2 Kb4 57.Bf6 Kc5 58.Ke3 Bc2 59.Bc3 Kb5 60.Kd4 Be4 61.Ke3 Ka4 62.Kd2 Bb1 63.Bd4 Kb3 64.Bf6 Kc4 65.Ke3 Kb5 66.Kd2 Kc5 67.Ke3 Be4
39/47 00:10 42,227k 4,132k +1.92 47. ... Be4 48.Kf2 f5 49.g3 Kd6 50.Bd4 Kc6 51.Ke2 a3 52.Kd2 Kb5 53.Kc3 a2 54.Kd2 Kc4 55.Ke3 Bb1 56.Bf6 Kb4 57.Bd4 Kb5 58.Bf6 Be4 59.Be5 Kb4 60.Bf6 Kb3 61.Kd2 Bb1 62.Bd4 Kc4 63.Ke3 Be4 64.Bf6 Bf3 65.Be5 Bg2 66.Bf6 Kb3 67.Kd2
40/47 00:11 46,511k 4,143k +1.92 47. ... Be4 48.Kf2 f5 49.g3 Kd6 50.Bd4 Kc6 51.Ke2 a3 52.Kd2 Kb5 53.Kc3 a2 54.Kd2 Kc4 55.Ke3 Bb1 56.Bf6 Kb4 57.Bd4 Kb5 58.Bf6 Be4 59.Be5 Kb4 60.Bf6 Kb3 61.Kd2 Bb1 62.Bd4 Kc4 63.Ke3 Be4 64.Bf6 Bf3 65.Be5 Bh1 66.Bf6 Bg2 67.Bd4 Kb3 68.Kd2
41/47 00:11 49,201k 4,147k +1.92 47. ... Be4 48.Kf2 f5 49.g3 Kd6 50.Bd4 Kc6 51.Ke2 a3 52.Kd2 Kb5 53.Kc3 a2 54.Kd2 Kc4 55.Ke3 Bb1 56.Bf6 Kb4 57.Bd4 Kb5 58.Bf6 Be4 59.Be5 Kb4 60.Bf6 Kb3 61.Kd2 Bb1 62.Bd4 Kc4 63.Ke3 Bc2 64.Be5 Kb5 65.Kd2 Bb1 66.Bf6 Kc5 67.Ke3 Be4 68.Bd4+ Kc6 69.Bc3
42/47 00:16 67,312k 4,118k +1.92 47. ... Be4 48.Kf2 f5 49.g3 Kd6 50.Bd4 Kc6 51.Ke2 a3 52.Kd2 Kb5 53.Kc3 a2 54.Kd2 Kc4 55.Ke3 Bb1 56.Bf6 Kb4 57.Bd4 Kb5 58.Bf6 Be4 59.Be5 Kb4 60.Bf6 Kb3 61.Kd2 Kc4 62.Ke3 Bb1 63.Bd4 Bc2 64.Be5 Kb5 65.Kd2 Bb1 66.Bf6 Kb4 67.Bd4 Be4 68.Bf6 Kc5 69.Ke3
43/48+ 00:20 84,126k 4,137k +1.99 47. ... a3
43/48+ 00:21 89,837k 4,164k +2.06 47. ... a3
43/48+ 00:23 97,532k 4,201k +2.17 47. ... a3
43/48+ 00:26 112,538k 4,230k +2.33 47. ... a3
43/53+ 00:30 128,870k 4,207k +2.54 47. ... a3
43/53+ 00:35 149,703k 4,229k +2.82 47. ... a3
43/57+ 00:41 177,459k 4,237k +3.19 47. ... a3
43/57 00:42 180,676k 4,227k +3.19 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Bd4 f4 54.gxf4 Kxf4 55.Bb2 Kf5 56.Kf3 Bb1 57.Bg7 Be4+ 58.Ke3 Kg4 59.Bf6 Bb1 60.Kf2 Kf5 61.Bh8 Ke4 62.Bb2 Kf4 63.Ba1 Be4 64.Ke2 Kg3 65.Bf6 Bb1 66.Be5+ Kh3 67.Bf6 d4 68.Bxd4 Kxh4 69.Kf3 g5 70.Kg2 Be4+ 71.Kh2
44/57 00:44 186,579k 4,236k +3.19 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg2 63.Bb2 Be4 64.Bd4 Kh3 65.Kb2 Bb1 66.Bf6 Kg4 67.Kc1 Kg3 68.Kb2 Kf3 69.Bc3 Ke4 70.Ba5 Ke3 71.Bb6+ Ke4 72.Bg1
45/57 00:53 231,713k 4,304k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3
46/57 00:54 232,747k 4,306k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kd3 75.Bf6
47/58 00:54 235,004k 4,310k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kd3 75.Bf6 Kc2 76.Kg1
48/59 00:57 246,509k 4,310k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kd3 75.Bf6 Kc2 76.Kg1 Kb1
49/60 00:58 251,100k 4,316k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kd3 75.Bf6 Kc2 76.Kg1 Kb1 77.Bd4
50/60 01:00 260,447k 4,303k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kd3 75.Bf6 Kc2 76.Kg1 Kb1 77.Bd4
51/66 01:27 381,050k 4,332k +3.22 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ke3 g5 66.Kf2 Kh3 67.Be5 g4 68.Kf1 g3 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kf4 75.Bg7 Kf5 76.Bb2 Bd3+ 77.Kg1 Kg4
52/66 01:29 386,132k 4,332k +3.22 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ke3 g5 66.Kf2 Be4 67.Bf6 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Bb1 74.Kf1 Be4 75.Bg7 Kf4 76.Bd4 Bb1 77.Bb2 Bd3+ 78.Kg1 Be4
53/66 01:32 402,072k 4,344k +3.22 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ke3 g5 66.Kf2 Be4 67.Bf6 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Bb1 74.Kf1 Be4 75.Bg7 Kf4 76.Bd4 Bb1 77.Bb2 Bd3+ 78.Kg1 Be4
54/66+ 02:33 657,981k 4,293k +3.30 47. ... a3
54/66 03:29 899,681k 4,286k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Bb1 74.Kf1 Be4 75.Bg7 Ke3 76.Ba1 Kd3
55/66 03:47 976,574k 4,291k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf3 79.Ba1 Ke3
56/68 03:50 986,232k 4,288k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf5 79.Bd4 Kf4 80.Bf6
57/68 03:54 1,006,719k 4,290k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf5 79.Bd4 Kf4 80.Bf6 Ke3
58/69 03:59 1,026,371k 4,290k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf5 79.Bd4 Kf4 80.Bf6 Bd3+ 81.Kg2 Bc2
59/70 04:30 1,158,892k 4,291k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf5 79.Bd4 Kf4 80.Bf6 Bd3+ 81.Kg2 Bc2 82.Bb2
60/70 04:53 1,261,646k 4,294k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf5 79.Bd4 Kf4 80.Bf6 Bd3+ 81.Kg2 Bc2 82.Bb2
61/73 04:57 1,276,898k 4,293k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bd4 Kh4 77.Bb2 Bc2 78.Kg2 Kg4 79.Bd4 Be4+ 80.Kg1 Kf4 81.Ba1 Kf3 82.Kf1 Ke3 83.Be5
62/73 05:08 1,324,457k 4,295k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bd4 Kh4 77.Bb2 Bc2 78.Kg2 Kg4 79.Be5 Bb1 80.Kf1 Kh3 81.Kg1 Be4 82.Bb2 Kg4 83.Kf1
63/73 05:17 1,363,706k 4,296k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bd4 Kh4 77.Bb2 Bc2 78.Kg2 Kg4 79.Be5 Bb1 80.Kf1 Kh3 81.Kg1 Be4 82.Bb2 Kg4 83.Kf1 Kf4
64/75+ 12:38 3,297,267k 4,347k +3.44 47. ... a3
64/79+ 17:34 4,563,684k 4,330k +3.51 47. ... a3
64/79+ 23:32 6,148,745k 4,353k +3.62 47. ... a3
64/79+ 28:03 7,364,557k 4,375k +3.77 47. ... a3
64/79+ 32:11 8,563,352k 4,434k +3.98 47. ... a3
64/79+ 36:59 9,859,384k 4,443k +4.27 47. ... a3
64/83+ 44:34 11,923,945k 4,458k +4.64 47. ... a3
64/83+ 52:18 14,003,050k 4,461k +5.12 47. ... a3
64/84+ 1:02:05 16,202,796k 4,350k +5.75 47. ... a3
64/84+ 1:13:44 18,916,281k 4,276k +6.54 47. ... a3
as usual, SF sees something Dvoretsky and co have failed to see. :)

I can't believe SF behaves paranoid here; engines first.
Top
Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 12:41 pm

Re: Shirov's Bh3, Houdini solves it

Post by Lyudmil Tsvetkov »

Uri Blass wrote:
Uri Blass wrote:Stockfish still did not solve the fail high on my computer.
Here is the rest of the analysis so far.

I still do not believe that a3 is winning.

[d]8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47

Stockfish_16121212_x64_modern:
64/89+ 1:26:57 22,351,073k 4,284k +7.56 47. ... a3
64/89+ 1:42:46 26,391,720k 4,280k +8.85 47. ... a3
64/89+ 2:01:14 31,413,235k 4,318k +10.48 47. ... a3
64/89+ 2:33:25 39,395,426k 4,280k +12.54 47. ... a3
64/98+ 3:34:26 56,441,530k 4,387k +15.14 47. ... a3
64/98+ 4:30:27 72,766,076k 4,484k +18.39 47. ... a3
64/98+ 5:19:13 87,012,668k 4,543k +22.49 47. ... a3
64/98+ 6:50:49 113,834,255k 4,618k +27.62 47. ... a3
some more analysis(still did not get a final score and unfortunately I trust no score of stockfish to be a winning score based on bad experience)

I wonder what is the plan of stockfish after 47...a3 48.Kf2 at depth 64 but stockfish does not show it in the output.

So far there are 20 fail high and one fail low at depth 64 with no final score.
Here is the interesting part when stockfish's branching factor is more than 100.

[d]8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47

Stockfish_16121212_x64_modern:
63/73 05:17 1,363,706k 4,296k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bd4 Kh4 77.Bb2 Bc2 78.Kg2 Kg4 79.Be5 Bb1 80.Kf1 Kh3 81.Kg1 Be4 82.Bb2 Kg4 83.Kf1 Kf4
64/75+ 12:38 3,297,267k 4,347k +3.44 47. ... a3
64/79+ 17:34 4,563,684k 4,330k +3.51 47. ... a3
64/79+ 23:32 6,148,745k 4,353k +3.62 47. ... a3
64/79+ 28:03 7,364,557k 4,375k +3.77 47. ... a3
64/79+ 32:11 8,563,352k 4,434k +3.98 47. ... a3
64/79+ 36:59 9,859,384k 4,443k +4.27 47. ... a3
64/83+ 44:34 11,923,945k 4,458k +4.64 47. ... a3
64/83+ 52:18 14,003,050k 4,461k +5.12 47. ... a3
64/84+ 1:02:05 16,202,796k 4,350k +5.75 47. ... a3
64/84+ 1:13:44 18,916,281k 4,276k +6.54 47. ... a3
64/89+ 1:26:57 22,351,073k 4,284k +7.56 47. ... a3
64/89+ 1:42:46 26,391,720k 4,280k +8.85 47. ... a3
64/89+ 2:01:14 31,413,235k 4,318k +10.48 47. ... a3
64/89+ 2:33:25 39,395,426k 4,280k +12.54 47. ... a3
64/98+ 3:34:26 56,441,530k 4,387k +15.14 47. ... a3
64/98+ 4:30:27 72,766,076k 4,484k +18.39 47. ... a3
64/98+ 5:19:13 87,012,668k 4,543k +22.49 47. ... a3
64/98+ 6:50:49 113,834,255k 4,618k +27.62 47. ... a3
64/98- 7:50:38 131,168,822k 4,645k +22.46 47. ... a3 48.Kf2
64/98+ 7:51:16 131,356,450k 4,645k +28.70 47. ... a3
64/100+ 9:10:51 153,001,076k 4,629k +38.81 47. ... a3
how can the branching factor be more than 100, when there are less than 100 legal moves in any position of this simple endgame?
Top
Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 12:41 pm

Re: Shirov's Bh3, Houdini solves it

Post by Lyudmil Tsvetkov »

MikeB wrote:
Uri Blass wrote:Stockfish still did not solve the fail high on my computer.
Here is the rest of the analysis so far.

I still do not believe that a3 is winning.

[d]8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47

Stockfish_16121212_x64_modern:
64/89+ 1:26:57 22,351,073k 4,284k +7.56 47. ... a3
64/89+ 1:42:46 26,391,720k 4,280k +8.85 47. ... a3
64/89+ 2:01:14 31,413,235k 4,318k +10.48 47. ... a3
64/89+ 2:33:25 39,395,426k 4,280k +12.54 47. ... a3
64/98+ 3:34:26 56,441,530k 4,387k +15.14 47. ... a3
64/98+ 4:30:27 72,766,076k 4,484k +18.39 47. ... a3
64/98+ 5:19:13 87,012,668k 4,543k +22.49 47. ... a3
64/98+ 6:50:49 113,834,255k 4,618k +27.62 47. ... a3
I'm not convinced on a3 either, so i thought I would play it out at 5 m 5s with 12 cores and using 18 threads. But this latest version finds Bh3 and then sees mate after a few moves.

[pgn][Event "Computer Chess Game"]
[Site "Mac-Pro.local"]
[Date "2016.12.16"]
[Round "-"]
[White "MacCfish-mf v8.1 64 POPC"]
[Black "MacCfish-mf v8.1 64 POPC"]
[Result "0-1"]
[TimeControl "300+5"]
[FEN "8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 1"]
[SetUp "1"]


1... Bh3 {+6.52/40 1:14} 2. Kf2 {-4.85/43 1:14} Kf5 {+9.13/24 10} 3. g3
{-4.87/25 7} Ke4 {+16.44/31 4} 4. Bb2 {-63.75/36 1:05} Bg4 {+56.22/32 5} 5.
Ke1 {-1000.33/40 17} Kd3 {+1000.38/38 15} 6. Kf1 {-1000.27/44 12}
0-1[/pgn]
you screwed up the engine entirely. :)
Top
Uri Blass
Posts: 11107
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Shirov's Bh3, Houdini solves it

Post by Uri Blass »

Uri Blass wrote:
Uri Blass wrote:Stockfish still did not solve the fail high on my computer.
Here is the rest of the analysis so far.

I still do not believe that a3 is winning.

[d]8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47

Stockfish_16121212_x64_modern:
64/89+ 1:26:57 22,351,073k 4,284k +7.56 47. ... a3
64/89+ 1:42:46 26,391,720k 4,280k +8.85 47. ... a3
64/89+ 2:01:14 31,413,235k 4,318k +10.48 47. ... a3
64/89+ 2:33:25 39,395,426k 4,280k +12.54 47. ... a3
64/98+ 3:34:26 56,441,530k 4,387k +15.14 47. ... a3
64/98+ 4:30:27 72,766,076k 4,484k +18.39 47. ... a3
64/98+ 5:19:13 87,012,668k 4,543k +22.49 47. ... a3
64/98+ 6:50:49 113,834,255k 4,618k +27.62 47. ... a3
some more analysis(still did not get a final score and unfortunately I trust no score of stockfish to be a winning score based on bad experience)

I wonder what is the plan of stockfish after 47...a3 48.Kf2 at depth 64 but stockfish does not show it in the output.

So far there are 20 fail high and one fail low at depth 64 with no final score.
Here is the interesting part when stockfish's branching factor is more than 100.

[d]8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47

Stockfish_16121212_x64_modern:
63/73 05:17 1,363,706k 4,296k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bd4 Kh4 77.Bb2 Bc2 78.Kg2 Kg4 79.Be5 Bb1 80.Kf1 Kh3 81.Kg1 Be4 82.Bb2 Kg4 83.Kf1 Kf4
64/75+ 12:38 3,297,267k 4,347k +3.44 47. ... a3
64/79+ 17:34 4,563,684k 4,330k +3.51 47. ... a3
64/79+ 23:32 6,148,745k 4,353k +3.62 47. ... a3
64/79+ 28:03 7,364,557k 4,375k +3.77 47. ... a3
64/79+ 32:11 8,563,352k 4,434k +3.98 47. ... a3
64/79+ 36:59 9,859,384k 4,443k +4.27 47. ... a3
64/83+ 44:34 11,923,945k 4,458k +4.64 47. ... a3
64/83+ 52:18 14,003,050k 4,461k +5.12 47. ... a3
64/84+ 1:02:05 16,202,796k 4,350k +5.75 47. ... a3
64/84+ 1:13:44 18,916,281k 4,276k +6.54 47. ... a3
64/89+ 1:26:57 22,351,073k 4,284k +7.56 47. ... a3
64/89+ 1:42:46 26,391,720k 4,280k +8.85 47. ... a3
64/89+ 2:01:14 31,413,235k 4,318k +10.48 47. ... a3
64/89+ 2:33:25 39,395,426k 4,280k +12.54 47. ... a3
64/98+ 3:34:26 56,441,530k 4,387k +15.14 47. ... a3
64/98+ 4:30:27 72,766,076k 4,484k +18.39 47. ... a3
64/98+ 5:19:13 87,012,668k 4,543k +22.49 47. ... a3
64/98+ 6:50:49 113,834,255k 4,618k +27.62 47. ... a3
64/98- 7:50:38 131,168,822k 4,645k +22.46 47. ... a3 48.Kf2
64/98+ 7:51:16 131,356,450k 4,645k +28.70 47. ... a3
64/100+ 9:10:51 153,001,076k 4,629k +38.81 47. ... a3
64/103- 10:40:43 178,414,161k 4,641k +25.13 47. ... a3 48.Kf2
64/103+ 10:45:23 179,786,553k 4,643k +40.81 47. ... a3
64/104 13:56:08 237,480,173k 4,734k +53.90 47. ... a3 48.Kf2 Bc2 49.g3 Kf5 50.Ke3 Be4 51.Kd4 Kg4 52.Ke3 Kxg3 53.Bxf6 a2 54.Ke2 d4 55.Bxd4 Kxh4 56.Kf1 g5 57.Ke2 Bc6 58.Kd1 Kh3 59.Bf6 g4 60.Bd4 g3 61.Kc2 Be4+ 62.Kc1 Bg2 63.Kc2 Bd5 64.Kc1 g2 65.Kb2 Kh2 66.Be3 g1Q 67.Bxg1+ Kxg1 68.Ka1 Kf2 69.Kb2 Ke3 70.Ka1 Bg8 71.Kb2 Kf3 72.Ka1 Bc4 73.Kb2 Kg3 74.Ka1 Bd5 75.Kb2 Kh3 76.Ka1 Kg4 77.Kb2 Kh4 78.Ka1 Kh3 79.Kb2 Kg4 80.Ka1 Bc4 81.Kb2 Kf3 82.Ka1 Kf2 83.Kb2 Kg1 84.Ka1 Kh2 85.Kb2 Kg2 86.Ka1 Kg1 87.Kb2 Bg8 88.Ka1 Bd5 89.Kb2


Finally got a line with score of 53.9 pawns for black.

stockfish does not suggest 49.Ke3 but 49.g3 in this line but it seems to me that the main line is a draw because all the d g h pawns fall and the king block the a pawn so we have a blind bishop that stockfish does not know to evaluate correctly and think that the blind bishop worth 53.9 pawns.
Top
peter
Posts: 3477
Joined: Sat Feb 16, 2008 7:38 am
Full name: Peter Martan

Re: Shirov's Bh3, Houdini solves it

Post by peter »

Uri Blass wrote: 64/104 13:56:08 237,480,173k 4,734k +53.90 47. ... a3 48.Kf2 Bc2 49.g3 Kf5 50.Ke3 Be4 51.Kd4 Kg4 52.Ke3 Kxg3 53.Bxf6 a2 54.Ke2 d4 55.Bxd4 Kxh4 56.Kf1 g5 57.Ke2 Bc6 58.Kd1 Kh3 59.Bf6 g4 60.Bd4 g3 61.Kc2 Be4+ 62.Kc1 Bg2 63.Kc2 Bd5 64.Kc1 g2 65.Kb2 Kh2 66.Be3 g1Q 67.Bxg1+ Kxg1 68.Ka1 Kf2 69.Kb2 Ke3 70.Ka1 Bg8 71.Kb2 Kf3 72.Ka1 Bc4 73.Kb2 Kg3 74.Ka1 Bd5 75.Kb2 Kh3 76.Ka1 Kg4 77.Kb2 Kh4 78.Ka1 Kh3 79.Kb2 Kg4 80.Ka1 Bc4 81.Kb2 Kf3 82.Ka1 Kf2 83.Kb2 Kg1 84.Ka1 Kh2 85.Kb2 Kg2 86.Ka1 Kg1 87.Kb2 Bg8 88.Ka1 Bd5 89.Kb2


Finally got a line with score of 53.9 pawns for black.

stockfish does not suggest 49.Ke3 but 49.g3 in this line but it seems to me that the main line is a draw because all the d g h pawns fall and the king block the a pawn so we have a blind bishop that stockfish does not know to evaluate correctly and think that the blind bishop worth 53.9 pawns.
This one Output- line loses finally (again not to speak about earlier moves) by 51.Kd4??, SF simply doesn't get the point, that White most not give away the g-Pawn without getting the f- Pawn at least the very next move, maybe even before that. Here e.g. with
51. Kf2 a2 52. Bb2 Kg4 53. Bxf6 draw.

You're right about the blindfold Bishop, but I had hoped for the Syzygys to equalize this blind spot.
Probably it doesn't here, because White can delay the 6some positions three times with Pawn- and capture-moves before the 50moves rule comes into horizon.

Edit: You didn't use Syzygys here, did you? I guess you yet should do so, as for Output- lines and Evals, e.g. doesn't SF show these bad blunders like in your SF's lines, here with me and full Syzgyg- usage,
Last edited by peter on Sat Dec 17, 2016 10:20 am, edited 1 time in total.
Peter.
Top
Uri Blass
Posts: 11107
Joined: Thu Mar 09, 2006 12:37 am
Location: Tel-Aviv Israel

Re: Shirov's Bh3, Houdini solves it

Post by Uri Blass »

Lyudmil Tsvetkov wrote:
Uri Blass wrote:
Lyudmil Tsvetkov wrote:
peter wrote:Dead draw, Lyudmil.

BTW, don't you think that to be a fittle funny too:
...do not look at what engines suggest...
...I ran a 10 games shootout with Komodo 10.1...
...games were at 1 min. bullet...
:)
I do not see what is so funny; you migth play 100 instead of 10 games, and black will still score around 30% wins or so. That is how engines are tuned, using stats of winning probabilities.

objectively migth be drawn (which I am still not certain), but black has nice winning chances after Kf5, capturing the h pawn and going to the first rank on f1.

problem however is not if this is a draw or win, but, as John wanted to know, why top engines are unable to find the key move. I myself do not see how an accurate eval assessment of this position could be returned without specifying some 500 lines of code, because:

- most eg positions with 2 passers and 3 files apart migth be drawn, but with c & f files wins
- positions with pawn more than 3 files apart might be wins, but with edge a or h pawns and wrong bishop draws; still, even some of those are wins
- many things will also depend on the placement of remaining pawns, some positions are draws, while others wins

so, if engines give +200cps for most of those, how do you make the distinction?
The problem is not that the engine has an evaluation that is not accurate.

It is ok not to know if the position is a win for black and to think 30% win for black and 70% draw but in this case I could expect an evaluation of something like +0.6 for black.

The main problem is that the engine claim that it is almost sure black is winning based on the evaluation when it suggest a drawing move.

I think that it is bad that people use tablebases because tablebases hide the bugs of chess engines so sometimes people even do not see how wrong they can be.

Latest version of stockfish from 12.12(without tablebases) show the following analysis and I wonder what cause it to fail high again and again on a3 and when it is going to be stopped(6.54 pawns at depth 64 after more than an hour but I do not believe that a3 is winning):

[d]8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47

Stockfish_16121212_x64_modern:
1/1 00:00 35 35k +1.79 47. ... a3
2/2 00:00 102 102k +2.17 47. ... a3 48.g3
3/3 00:00 177 177k +2.52 47. ... a3 48.Ba1 a2
4/4 00:00 262 262k +2.52 47. ... a3 48.Ba1 a2 49.g3
5/6 00:00 2k 788k +1.78 47. ... a3 48.Kf2 Be4 49.Bd4 a2
6/6 00:00 4k 1,343k +1.73 47. ... a3 48.g3 a2 49.Kf2 Bd3 50.Kf3 Be4+ 51.Ke2
7/7 00:00 5k 1,606k +1.70 47. ... a3 48.g3 a2 49.Kf2 Be4 50.Bd4 f5
8/8 00:00 8k 2,013k +1.64 47. ... a3 48.g3 Bc2 49.Kf2 a2 50.Bd4 f5 51.Ke3
9/11 00:00 25k 2,779k +1.48 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 a2 51.Bd4 Bb3 52.Kd3
10/12 00:00 34k 2,801k +1.45 47. ... a3 48.Kf2 Be4 49.g3 Bc2 50.Ke3 Bf5 51.Kd4 Bc2 52.Kc5 a2
11/15 00:00 64k 2,922k +1.47 47. ... a3 48.Kf2 Be4 49.g3 Kf5 50.Ba1 Bc2 51.Ke3 Bb3 52.Bd4 Bc4 53.Kd2 a2
12/16 00:00 101k 2,980k +1.42 47. ... a3 48.Kf2 Bc2 49.Ke2 Kf5 50.Kd2 Bb1 51.g3 Kg4 52.Bxf6 Kxg3 53.Kc1 Be4
13/16 00:00 120k 3,073k +1.41 47. ... a3 48.Kf2 Bd3 49.g3 Kf5 50.Ke3 Bb1 51.Ke2 Kg4 52.Bxf6 Kxg3 53.Kd2 Kf4 54.Kc1
14/16 00:00 122k 3,049k +1.41 47. ... a3 48.Kf2 Bc2 49.g3 Kf5 50.Ke3 Bb1 51.Ke2 Kg4 52.Bxf6 Kxg3 53.Kd2 Kf4 54.Kc1
15/23 00:00 228k 3,504k +1.41 47. ... a3 48.Kf2 Be4 49.g3 f5 50.Ke3 a2 51.Kd4 Kd6 52.Bb2 Bf3 53.Kc3 Be2 54.Kb3 Bc4+ 55.Kc2
16/23 00:00 250k 3,577k +1.43 47. ... a3 48.Kf2 Be4 49.g3 f5 50.Ke3 Kd7 51.Kd2 Kc6 52.Bf6 Kb5 53.Kc3 Ka4 54.Bg7 Bh1 55.Kc2 Bg2 56.Bd4 Kb4
17/24 00:00 619k 3,558k +1.68 47. ... Be4 48.g4 a3 49.h5 gxh5 50.gxh5 f5 51.Kf2 f4 52.Bd4 a2 53.Ke1 Kd6 54.Kd1 f3 55.Ke1 Kc6 56.h6 Kb5
18/24 00:00 669k 3,503k +1.69 47. ... Be4 48.g4 a3 49.Kf2 f5 50.gxf5+ Kxf5 51.Ke3 Kg4 52.Bf6 a2 53.Kd2 Bb1 54.Bc3 Kxh4 55.Kc1 g5 56.Bf6 Kh3 57.Be5 g4
19/29 00:00 919k 3,283k +1.80 47. ... Be4 48.g4 a3 49.Kf2 f5 50.gxf5+ Kxf5 51.Ke3 a2 52.Bd4 Kg4 53.Kd2 Kxh4 54.Kc1 g5 55.Kb2 g4 56.Kxa2 g3 57.Kb3 g2 58.Kb4 Kg4
20/31 00:00 1,356k 3,095k +2.04 47. ... Be4 48.g4 a3 49.h5 gxh5 50.gxh5 f5 51.Kf2 f4 52.Ke2 Kf5 53.h6 Kg5 54.Bg7 a2 55.Kf2 Kg4 56.Ke2 Bg6 57.Bd4 Kf5 58.Kf3 Kg5 59.h7 Be4+ 60.Ke2 Bxh7
21/33 00:00 2,072k 3,465k +2.14 47. ... Be4 48.g4 a3 49.h5 gxh5 50.gxh5 f5 51.Kf2 f4 52.h6 a2 53.Ke2 Bg6 54.Kf3 Kf5 55.Bd4 Kg5 56.Bg7 Be4+ 57.Ke2 Kh5 58.Kf2 Kg4 59.Bd4 Bg6 60.Bg7 f3 61.Ke3 Kg3 62.Be5+ Kg2
22/34 00:00 2,532k 3,617k +2.07 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Ba1 a2 56.Bf6 Bf5 57.Bd4 Bh3 58.Bf6 Bd7 59.Ba1
23/34 00:00 3,076k 3,715k +2.00 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Bf6 a2 56.Ba1 Bf5 57.Bd4 Kb3 58.Kd2 Bd7 59.Ba1 Kc4 60.Ke3 Bg4 61.Bf6
24/34 00:00 3,609k 3,815k +1.93 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 a2 56.Bf6 Bf5 57.Bd4 Bd7 58.Ba1 Bh3 59.Bf6 Bg2 60.Bd4 Bh3 61.Ba1 Bc8
25/34 00:01 3,862k 3,847k +1.93 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 a2 56.Bf6 Bf5 57.Bd4 Bd7 58.Ba1 Bh3 59.Bf6 Kb4 60.Kd2 Be6 61.Kc1 Kb3 62.Ba1
26/35 00:01 4,237k 3,894k +2.00 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bd4 Kb3 57.Kd2 Bf3 58.Ke3 Be4 59.Kd2 Kc4 60.Ke3 a2 61.Bf6 Bf5 62.Bd4 Bd7 63.Bf6
27/35 00:01 5,252k 3,817k +1.92 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 a2 56.Bf6 Bf5 57.Bd4 Bh3 58.Bf6 Kc5 59.Bc3 Be6 60.Bd4+ Kc4 61.Bf6 Kb4 62.Ba1
28/35 00:01 5,968k 3,765k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 Kb4 57.Bd4 Kb3 58.Kd2 Bf3 59.Ke3 Be4 60.Kd2 Kc4 61.Ke3 Kb4 62.Be5 a2 63.Kd2 Kc4 64.Kc1 d4
29/35 00:01 6,922k 3,793k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 a2 57.Bd4 Be4 58.Be5 Bb1 59.Bf6 Bf5 60.Be5 Bd7 61.Bf6 Be6 62.Be5 Bg4 63.Bf6 Bh3 64.Ba1 Kb3
30/35 00:02 7,987k 3,853k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 a2 57.Bd4 Be4 58.Be5 Bb1 59.Bf6 Bf5 60.Be5 Bd7 61.Bf6 Be6 62.Be5 Bg4 63.Bf6 Bh3 64.Ba1 Kb3
31/37 00:02 9,220k 3,802k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 a2 57.Bd4 Be4 58.Be5 Bb1 59.Bf6 Bf5 60.Be5 Bd7 61.Bf6 Be6 62.Be5 Bg4 63.Bf6 Kc5 64.Bd4+ Kb4 65.Ba1 Kc4
32/40 00:02 9,938k 3,793k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 Bh3 57.Bd4 Kb3 58.Kd2 Bg4 59.Bf6 Kc4 60.Ke3 Be6 61.Bd4 Bf5 62.Be5 Be4 63.Bd4 Bh1 64.Bf6 Bg2 65.Bd4 Kb3 66.Kd2
33/40 00:03 12,830k 3,837k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 Bh1 57.Bd4 Kb3 58.Kd2 Bf3 59.Bf6 Kc4 60.Ke3 Bg4 61.Ba1 Bd7 62.Bd4 Bh3 63.Be5 Be6 64.Bf6 Bf5 65.Be5 Bg4
34/40 00:03 15,124k 3,877k +1.91 47. ... Be4 48.g4 a3 49.Kf2 f5 50.Kg3 fxg4 51.Kxg4 Kd7 52.Kf4 Kc6 53.Ke3 Kc5 54.Bd4+ Kc4 55.Be5 Bg2 56.Bf6 Bh1 57.Bd4 Kb3 58.Kd2 Bf3 59.Bf6 Kc4 60.Ke3 Bg4 61.Ba1 Bd7 62.Bd4 Bh3 63.Be5 Be6 64.Bf6 Bf5 65.Be5 Be4 66.Bf6
35/45 00:05 21,185k 3,895k +1.91 47. ... Be4 48.g4 f5 49.gxf5+ Bxf5 50.Kf2 Kd6 51.Bb4+ Kc6 52.Ke3 Kb5 53.Be7 Kc4 54.Kd2 Kb3 55.Bd6 a3 56.Be5 Kc4 57.Ke3 Bh3 58.Bd4 Be6 59.Ba1 Bd7 60.Bd4 Bh3 61.Be5 Kb3 62.Kd2 Bf5 63.Bf6 Kc4 64.Ke3 Bb1 65.Bd4 Be4 66.Be5 Bg2
36/47 00:06 25,725k 3,952k +1.91 47. ... Be4 48.g4 f5 49.gxf5+ Bxf5 50.Kf2 Kd6 51.Bb4+ Kc6 52.Ke3 Kb5 53.Be7 Kc4 54.Kd2 Kb3 55.Bd6 a3 56.Be5 Kc4 57.Ke3 Bh3 58.Bd4 Be6 59.Be5 Bd7 60.Bd4 Bh3 61.Be5 Kb3 62.Kd2 Bf5 63.Bf6 Kc4 64.Ke3 Bb1 65.Bd4 Be4 66.Be5 Bg2 67.Bf6 Bh1 68.Bd4 Kb3 69.Kd2
37/47 00:07 28,701k 4,002k +1.91 47. ... Be4 48.g4 f5 49.gxf5+ Bxf5 50.Kf2 Kd6 51.Bb4+ Kc6 52.Ke3 Kb5 53.Be7 Kc4 54.Kd2 Kb3 55.Bd6 a3 56.Be5 Kc4 57.Ke3 Bh3 58.Bd4 Be6 59.Be5 Bd7 60.Bd4 Bh3 61.Be5 Kb3 62.Kd2 Bf5 63.Bf6 Kc4 64.Ke3 Bb1 65.Bd4 Be4 66.Be5 Bg2 67.Bf6 Bh1 68.Bd4 Kb3 69.Kd2 Bf3
38/47+ 00:08 34,047k 4,048k +1.99 47. ... Be4
38/47 00:10 41,473k 4,127k +1.92 47. ... Be4 48.Kf2 f5 49.g3 Kd6 50.Bd4 Kc6 51.Ke2 a3 52.Kd2 a2 53.Kc1 Kb5 54.Kd2 Kc4 55.Ke3 Kb3 56.Kd2 Kb4 57.Bf6 Kc5 58.Ke3 Bc2 59.Bc3 Kb5 60.Kd4 Be4 61.Ke3 Ka4 62.Kd2 Bb1 63.Bd4 Kb3 64.Bf6 Kc4 65.Ke3 Kb5 66.Kd2 Kc5 67.Ke3 Be4
39/47 00:10 42,227k 4,132k +1.92 47. ... Be4 48.Kf2 f5 49.g3 Kd6 50.Bd4 Kc6 51.Ke2 a3 52.Kd2 Kb5 53.Kc3 a2 54.Kd2 Kc4 55.Ke3 Bb1 56.Bf6 Kb4 57.Bd4 Kb5 58.Bf6 Be4 59.Be5 Kb4 60.Bf6 Kb3 61.Kd2 Bb1 62.Bd4 Kc4 63.Ke3 Be4 64.Bf6 Bf3 65.Be5 Bg2 66.Bf6 Kb3 67.Kd2
40/47 00:11 46,511k 4,143k +1.92 47. ... Be4 48.Kf2 f5 49.g3 Kd6 50.Bd4 Kc6 51.Ke2 a3 52.Kd2 Kb5 53.Kc3 a2 54.Kd2 Kc4 55.Ke3 Bb1 56.Bf6 Kb4 57.Bd4 Kb5 58.Bf6 Be4 59.Be5 Kb4 60.Bf6 Kb3 61.Kd2 Bb1 62.Bd4 Kc4 63.Ke3 Be4 64.Bf6 Bf3 65.Be5 Bh1 66.Bf6 Bg2 67.Bd4 Kb3 68.Kd2
41/47 00:11 49,201k 4,147k +1.92 47. ... Be4 48.Kf2 f5 49.g3 Kd6 50.Bd4 Kc6 51.Ke2 a3 52.Kd2 Kb5 53.Kc3 a2 54.Kd2 Kc4 55.Ke3 Bb1 56.Bf6 Kb4 57.Bd4 Kb5 58.Bf6 Be4 59.Be5 Kb4 60.Bf6 Kb3 61.Kd2 Bb1 62.Bd4 Kc4 63.Ke3 Bc2 64.Be5 Kb5 65.Kd2 Bb1 66.Bf6 Kc5 67.Ke3 Be4 68.Bd4+ Kc6 69.Bc3
42/47 00:16 67,312k 4,118k +1.92 47. ... Be4 48.Kf2 f5 49.g3 Kd6 50.Bd4 Kc6 51.Ke2 a3 52.Kd2 Kb5 53.Kc3 a2 54.Kd2 Kc4 55.Ke3 Bb1 56.Bf6 Kb4 57.Bd4 Kb5 58.Bf6 Be4 59.Be5 Kb4 60.Bf6 Kb3 61.Kd2 Kc4 62.Ke3 Bb1 63.Bd4 Bc2 64.Be5 Kb5 65.Kd2 Bb1 66.Bf6 Kb4 67.Bd4 Be4 68.Bf6 Kc5 69.Ke3
43/48+ 00:20 84,126k 4,137k +1.99 47. ... a3
43/48+ 00:21 89,837k 4,164k +2.06 47. ... a3
43/48+ 00:23 97,532k 4,201k +2.17 47. ... a3
43/48+ 00:26 112,538k 4,230k +2.33 47. ... a3
43/53+ 00:30 128,870k 4,207k +2.54 47. ... a3
43/53+ 00:35 149,703k 4,229k +2.82 47. ... a3
43/57+ 00:41 177,459k 4,237k +3.19 47. ... a3
43/57 00:42 180,676k 4,227k +3.19 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Bd4 f4 54.gxf4 Kxf4 55.Bb2 Kf5 56.Kf3 Bb1 57.Bg7 Be4+ 58.Ke3 Kg4 59.Bf6 Bb1 60.Kf2 Kf5 61.Bh8 Ke4 62.Bb2 Kf4 63.Ba1 Be4 64.Ke2 Kg3 65.Bf6 Bb1 66.Be5+ Kh3 67.Bf6 d4 68.Bxd4 Kxh4 69.Kf3 g5 70.Kg2 Be4+ 71.Kh2
44/57 00:44 186,579k 4,236k +3.19 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg2 63.Bb2 Be4 64.Bd4 Kh3 65.Kb2 Bb1 66.Bf6 Kg4 67.Kc1 Kg3 68.Kb2 Kf3 69.Bc3 Ke4 70.Ba5 Ke3 71.Bb6+ Ke4 72.Bg1
45/57 00:53 231,713k 4,304k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3
46/57 00:54 232,747k 4,306k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kd3 75.Bf6
47/58 00:54 235,004k 4,310k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kd3 75.Bf6 Kc2 76.Kg1
48/59 00:57 246,509k 4,310k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kd3 75.Bf6 Kc2 76.Kg1 Kb1
49/60 00:58 251,100k 4,316k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kd3 75.Bf6 Kc2 76.Kg1 Kb1 77.Bd4
50/60 01:00 260,447k 4,303k +3.20 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ba1 g5 66.Ke2 g4 67.Kf2 Kh3 68.Be5 g3+ 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kd3 75.Bf6 Kc2 76.Kg1 Kb1 77.Bd4
51/66 01:27 381,050k 4,332k +3.22 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ke3 g5 66.Kf2 Kh3 67.Be5 g4 68.Kf1 g3 69.Kg1 Bd3 70.Ba1 Kg4 71.Bd4 Be4 72.Kf1 Kf3 73.Bb2 Ke3 74.Ba1 Kf4 75.Bg7 Kf5 76.Bb2 Bd3+ 77.Kg1 Kg4
52/66 01:29 386,132k 4,332k +3.22 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ke3 g5 66.Kf2 Be4 67.Bf6 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Bb1 74.Kf1 Be4 75.Bg7 Kf4 76.Bd4 Bb1 77.Bb2 Bd3+ 78.Kg1 Be4
53/66 01:32 402,072k 4,344k +3.22 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd2 d4 64.Bxd4 Kxh4 65.Ke3 g5 66.Kf2 Be4 67.Bf6 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Bb1 74.Kf1 Be4 75.Bg7 Kf4 76.Bd4 Bb1 77.Bb2 Bd3+ 78.Kg1 Be4
54/66+ 02:33 657,981k 4,293k +3.30 47. ... a3
54/66 03:29 899,681k 4,286k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Bb1 74.Kf1 Be4 75.Bg7 Ke3 76.Ba1 Kd3
55/66 03:47 976,574k 4,291k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf3 79.Ba1 Ke3
56/68 03:50 986,232k 4,288k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf5 79.Bd4 Kf4 80.Bf6
57/68 03:54 1,006,719k 4,290k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf5 79.Bd4 Kf4 80.Bf6 Ke3
58/69 03:59 1,026,371k 4,290k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf5 79.Bd4 Kf4 80.Bf6 Bd3+ 81.Kg2 Bc2
59/70 04:30 1,158,892k 4,291k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf5 79.Bd4 Kf4 80.Bf6 Bd3+ 81.Kg2 Bc2 82.Bb2
60/70 04:53 1,261,646k 4,294k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bg7 Kf4 77.Kg1 Kg4 78.Kf1 Kf5 79.Bd4 Kf4 80.Bf6 Bd3+ 81.Kg2 Bc2 82.Bb2
61/73 04:57 1,276,898k 4,293k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bd4 Kh4 77.Bb2 Bc2 78.Kg2 Kg4 79.Bd4 Be4+ 80.Kg1 Kf4 81.Ba1 Kf3 82.Kf1 Ke3 83.Be5
62/73 05:08 1,324,457k 4,295k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bd4 Kh4 77.Bb2 Bc2 78.Kg2 Kg4 79.Be5 Bb1 80.Kf1 Kh3 81.Kg1 Be4 82.Bb2 Kg4 83.Kf1
63/73 05:17 1,363,706k 4,296k +3.37 47. ... a3 48.Kf2 Bc2 49.Ke3 Kf5 50.g3 Kg4 51.Kf2 f5 52.Be5 a2 53.Kg2 Be4+ 54.Kf2 Bb1 55.Kg2 Bc2 56.Kf2 Be4 57.Bf6 f4 58.gxf4 Kxf4 59.Ke2 Kg4 60.Ke3 Bb1 61.Kd2 Kh3 62.Kc1 Kg3 63.Kd1 d4 64.Bxd4 Kxh4 65.Ke1 g5 66.Bf6 Be4 67.Kf2 Kh3 68.Kg1 Kg3 69.Kf1 g4 70.Kg1 Kf3 71.Kf1 g3 72.Bd4 Bd3+ 73.Kg1 Kg4 74.Be5 Be4 75.Kf1 Kg5 76.Bd4 Kh4 77.Bb2 Bc2 78.Kg2 Kg4 79.Be5 Bb1 80.Kf1 Kh3 81.Kg1 Be4 82.Bb2 Kg4 83.Kf1 Kf4
64/75+ 12:38 3,297,267k 4,347k +3.44 47. ... a3
64/79+ 17:34 4,563,684k 4,330k +3.51 47. ... a3
64/79+ 23:32 6,148,745k 4,353k +3.62 47. ... a3
64/79+ 28:03 7,364,557k 4,375k +3.77 47. ... a3
64/79+ 32:11 8,563,352k 4,434k +3.98 47. ... a3
64/79+ 36:59 9,859,384k 4,443k +4.27 47. ... a3
64/83+ 44:34 11,923,945k 4,458k +4.64 47. ... a3
64/83+ 52:18 14,003,050k 4,461k +5.12 47. ... a3
64/84+ 1:02:05 16,202,796k 4,350k +5.75 47. ... a3
64/84+ 1:13:44 18,916,281k 4,276k +6.54 47. ... a3
as usual, SF sees something Dvoretsky and co have failed to see. :)

I can't believe SF behaves paranoid here; engines first.
Unfortunately I do not believe stockfish because I know that big scores of it can be wrong from experience and I suspect something wrong about pruning.

I think that tablebases only hide the pruning problem of stockfish in this case so it is better not to use them if you are interested in stockfish's bugs that I believe that are not only about tablebases positions.

Stockfish is basically an engine that you cannot trust about evaluations that it shows because the developers do not care about it but only about elo in bullet games.

Even 53.90 pawns by stockfish is not a win as this example show
and I believe that it is possible to find a simpler example that it happens with 1 core and a few seconds.

I think that it is not enough to have a good evaluation for special endgames and you need also special pruning rules for them.

For example do not search what you know is a draw(even if you are not in tablebases).

Here stockfish should stop to search even without tablebases as soon as it see a position of King and bishop and King and bishop and 2 pawns when it is obvious that white can sacrifice the bishop for a pawn so black remain with the blind bishop(this rule is not only about tablebases and it can be used to know that one side is not losing for example if you have KBPP vs KBPP when it is obvious that one side can sacrifice the bishop for a pawn and leave the opponent with the blind bishop and wrong pawn against 2 pawns when the wrong pawn cannot go to another file by capture that may be possible in some position like h vs g pawn).
Top
yanquis1972
Posts: 1766
Joined: Wed Jun 03, 2009 12:14 am

Re: Shirov's Bh3, Houdini solves it

Post by yanquis1972 »

good points & a good case youre making uri.

its a free engine, everyone else is playing catch up, & still its just chasing STC elo instead of becoming a well-rounded analysis tool. theres no money to be gained from topping elo lists AND it'd be on top of them again quickly anyway. id def like to see either a fork or, even better, the SF team to adjust their sights...
Top
Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 12:41 pm

Re: Shirov's Bh3, Houdini solves it

Post by Lyudmil Tsvetkov »

yanquis1972 wrote:good points & a good case youre making uri.

its a free engine, everyone else is playing catch up, & still its just chasing STC elo instead of becoming a well-rounded analysis tool. theres no money to be gained from topping elo lists AND it'd be on top of them again quickly anyway. id def like to see either a fork or, even better, the SF team to adjust their sights...
unfortunately not that easy; you can change an engine, add many useful terms that should improve play in general, but how to tune them?

especially in an engine that is emphatically formulistically tuned?

and they do not want to go out of the way to implement new knowledge, just changing a single formula, etc. for example, they try to implement chains, they just change a single formula and, of course, nothing works. it can not work, because 1st: in order to implement chains, you should first decouple the bonus for connected defended pawns and connected pawns on the same rank next to each other; and 2nd: decrease the general defended pawn bonus, to only after that administer the new chain bonus, both are somewhat redundant, you know.

etc., etc., so, it is not clear to me how to tune terms even if you know your engine by heart, so tuning when you do not know all SF code by heart is seemingly even more challenging.
Top
tpoppins
Posts: 919
Joined: Tue Nov 24, 2015 9:11 pm
Location: upstate

Re: Shirov's Bh3, Houdini solves it

Post by tpoppins »

From a post a few pages back:
tpoppins wrote:Topalov - Shirov, Linares 1998 (r10)
[d]8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47
Current Let's Check status:
Image
Three weeks later:

Image

Here are the goods:

[pgn][Event "Linares 15th"]
[Site "Linares"]
[Date "1998.03.04"]
[Round "10"]
[White "Topalov, Veselin"]
[Black "Shirov, Alexei"]
[Result "0-1"]
[ECO "D85"]
[WhiteElo "2740"]
[BlackElo "2710"]
[Annotator "SF 121216 x64 POPCNT"]
[SetUp "1"]
[FEN "8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47"]
[PlyCount "63"]


47... Bh3 $3 48. Kh2 ({after} 48. Kf2 {
Black mates a move quicker:} Kf5 49. g4+ Bxg4 50. Ke3 a3 51. Ba1 a2 52. Bb2
Bh3 53. Kd4 Bg2 54. Kc5 g5 55. h5 Be4 56. Kb4 g4 57. h6 Kg6 58. h7 Kxh7 59. Kb3
Bb1 60. Bc3 Kh6 61. Kb2 Kg5 62. Bd2+ Kh4 63. Bh6 g3 64. Bg7 f5 65. Ka3 f4 66.
Ka4 Be4 67. Be5 f3 68. Kb4 Kh3 69. Kc5 f2 70. Kd6 f1=Q 71. Ke7 a1=Q 72. Bxa1
Qxa1 73. Ke6 Qa6+ 74. Kd7 g2 75. Ke8 g1=Q 76. Kd8 Qg7 77. Ke8 Qa8#) 48... Bg4
49. Kg3 Kf5 50. Bb2 Bh5 51. Kf2 Ke4 52. Ke1 Ke3 53. g3 f5 54. Kf1 Kd2 55. Kf2
Bg4 56. Kf1 Kc2 57. Bd4 a3 58. Bc5 a2 59. Bd4 Kb1 60. Kf2 a1=Q 61. Bxa1 Kxa1
62. Ke3 Kb2 63. Kd4 Bf3 64. Kd3 Kb3 65. h5 gxh5 66. Kd4 Be4 67. Ke5 Kc3 68. Kf6
d4 69. Ke5 d3 70. Kd6 d2 71. Ke7 d1=Q 72. Ke6 Qd8 73. g4 hxg4 74. Ke5 Qe7+ 75.
Kf4 g3 76. Kxg3 Qg5+ 77. Kh3 Qg2+ 78. Kh4 Qg4# 0-1
[/pgn]

Backsolved with SF 121216 x64 POPCNT and 16 GB hash over the course of a sleepless December night, TBs off; verified with same without clearing the hash:

[d]8/8/4kpp1/3p4/p6P/2B4b/6P1/6K1 w - - 0 48

Code: Select all

Analysis by Stockfish 121216 x64 POPCNT:

1. -+ (-#30): 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Kd3 55.Bb2 Bg4 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.Ke5 d3 66.h5 gxh5 67.Kxf5 d2 68.Kf6 d1Q 69.Ke7 h4 70.Kf7 Qf1+ 71.Ke8 Qxg2 72.Ke7 h3 73.Kf7 h2 74.Kf6 Qc6+ 75.Kf5 h1Q 76.Kg5 Qhf3 77.Kh4 Qcf6# 
2. -+ (-#29): 48.Kf2 Kf5 49.Bb2 Ke4 50.g3 Bg4 51.Ke1 Kd3 52.Kf1 Kc2 53.Bxf6 a3 54.Be7 d4 55.Bxa3 d3 56.Bb4 d2 57.Bxd2 Kxd2 58.Kf2 Kd3 59.Kf1 Ke3 60.Kg2 Ke2 61.Kg1 Kf3 62.Kf1 Kxg3 63.Ke1 Kxh4 64.Kf2 g5 65.Kg2 Bh5 66.Kf2 g4 67.Ke2 g3+ 68.Kf1 Bf3 69.Ke1 g2 70.Kf2 Kg4 71.Kg1 Kf4 72.Kh2 Ke4 73.Kg1 Ke3 74.Kh2 Kf2 75.Kh3 g1Q 76.Kh4 Qg3# 
  Depth: 76/80   05:12:31  752255 MN

And a single-PV run at the root done the same way, starting with the iteration that shows a complete mating line. Quite a few of those here

8/8/4kpp1/3p1b2/p6P/2B5/6P1/6K1 b - - 0 47

Code: Select all

47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.Ke5 d3 66.h5 gxh5 67.Kxf5 d2 68.Kf6 d1Q 69.Ke7 Qe2+ 70.Kf7 Qxg2 71.Ke7 Qg6 72.Kd7 h4 73.Kd8 h3 74.Ke7 h2 75.Kd7 h1Q 76.Kc8 Qgc6+ 77.Kb8 Qb7# 
  -+ (-#31)  Depth: 56/66   00:00:01  23364kN

[identical PVs snipped]

47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.Ke5 d3 66.h5 gxh5 67.Kxf5 d2 68.Kf6 d1Q 69.Ke7 Qe2+ 70.Kf7 Qxg2 71.Ke7 Qg6 72.Kd7 h4 73.Kd8 h3 74.Ke7 h2 75.Kd7 h1Q 76.Kc8 Qgc6+ 77.Kb8 Qb7# 
  -+ (-#31)  Depth: 78/66   00:04:18  9975MN
47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.Ke5 d3 66.h5 gxh5 67.Kxf5 d2 68.Kf6 d1Q 69.Ke7 Qe2+ 70.Kf7 Qxg2 71.Ke7 Qd5 72.Kf8 h4 73.Ke7 h3 74.Kf6 h2 75.Ke7 h1Q 76.Kf6 Qh8+ 77.Kg6 Qdh5# 
  -+ (-#31)  Depth: 79/66   00:23:19  54323MN
47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.Ke5 d3 66.h5 gxh5 67.Kxf5 d2 68.Kf6 d1Q 69.Ke7 Qd5 70.g3 Kc3 71.Kf6 Kd4 72.g4 hxg4 73.Kg6 Ke5 74.Kg7 Qe6 75.Kh7 Kf6 76.Kh8 Qe7 77.Kg8 Qg7# 
  -+ (-#31)  Depth: 80/66   00:38:08  89404MN
47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.Ke5 d3 66.h5 gxh5 67.Kxf5 d2 68.Kf6 d1Q 69.Ke7 Qe2+ 70.Kf7 Qxg2 71.Ke7 Qd5 72.Kf8 Qd7 73.Kg8 Kc3 74.Kf8 Kd4 75.Kg8 Ke5 76.Kf8 Kf6 77.Kg8 Qg7# 
  -+ (-#31)  Depth: 81/66   02:07:53  302530MN

[funny how these two last PVs end with the same move without being identical]

47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.g4 Bxg4 66.Kxg4 d3 67.Kg5 d2 68.h5 gxh5 69.Kxh5 d1Q+ 70.Kg5 Qf1 71.Kg4 Kc3 72.Kg3 Qf6 73.Kg4 Kd4 74.Kh5 Ke3 75.Kg4 Qg6+ 76.Kh4 Kf3 77.Kh3 Qh6# 
  -+ (-#31)  Depth: 82/66   02:08:57  305097MN
47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.g4 Bxg4 66.Kxg4 d3 67.Kg5 d2 68.h5 gxh5 69.Kxh5 d1Q+ 70.Kg5 Qf1 71.Kg4 Kc3 72.Kg3 Qf6 73.Kg4 Kd4 74.Kh5 Ke3 75.Kg4 Qg6+ 76.Kh4 Kf3 77.Kh3 Qh6# 
  -+ (-#31)  Depth: 83/66   02:44:51  391300MN
47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.g4 Bxg4 66.Kxg4 d3 67.Kg5 d2 68.h5 gxh5 69.Kxh5 d1Q+ 70.Kg5 Qf1 71.Kg4 Qf7 72.Kg3 Kc3 73.Kg4 Kd4 74.Kh3 Qg7 75.Kh4 Ke4 76.Kh5 Kf4 77.Kh4 Qg4# 
  -+ (-#31)  Depth: 84/70   04:46:16  681270MN
47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.g4 Bxg4 66.Kxg4 d3 67.Kg5 d2 68.h5 gxh5 69.Kxh5 d1Q+ 70.Kg6 Qf1 71.Kg5 Kc3 72.Kg4 Kd4 73.Kg3 Ke4 74.Kh2 Qe2+ 75.Kg3 Kf5 76.Kh3 Kf4 77.Kh4 Qh2# 
  -+ (-#31)  Depth: 85/70   05:16:16  751516MN
47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.g4 Bxg4 66.Kxg4 d3 67.Kg5 d2 68.h5 gxh5 69.Kxh5 d1Q+ 70.Kg6 Qf1 71.Kg5 Kc3 72.Kg4 Kd4 73.Kg3 Ke4 74.Kh2 Qe2+ 75.Kg3 Kf5 76.Kh3 Kf4 77.Kh4 Qh2# 
  -+ (-#31)  Depth: 86/70   07:31:39  1077035MN
47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.g4 Bxg4 66.Kxg4 d3 67.Kg5 d2 68.h5 gxh5 69.Kxh5 d1Q+ 70.Kg6 Qf1 71.Kg5 Kc3 72.Kg4 Kd4 73.Kg3 Ke4 74.Kh2 Qe2+ 75.Kg3 Kf5 76.Kh3 Kf4 77.Kh4 Qh2# 
  -+ (-#31)  Depth: 87/70   09:32:52  1371210MN
47...Bh3 48.Kh2 Bg4 49.Bb2 Kf5 50.Kg3 Ke4 51.Kf2 f5 52.Kg3 Bh5 53.Bc1 f4+ 54.Kf2 Bg4 55.Bb2 Kd3 56.Ke1 Kc2 57.Be5 a3 58.Bd6 a2 59.Be5 Kb1 60.Kd2 Bf5 61.Ke2 a1Q 62.Bxa1 Kxa1 63.Kf3 d4 64.Kxf4 Kb2 65.g4 Bxg4 66.Kxg4 d3 67.Kg5 d2 68.h5 gxh5 69.Kxh5 d1Q+ 70.Kg6 Qf1 71.Kg5 Kc3 72.Kg4 Kd4 73.Kg3 Ke4 74.Kh2 Qe2+ 75.Kg3 Kf5 76.Kh3 Kf4 77.Kh4 Qh2# 
  -+ (-#31)  Depth: 88/70   13:45:18  1982702MN
Top

Return to “Computer Chess Club: General Topics”