mersenneforum.org

mersenneforum.org (https://www.mersenneforum.org/index.php)
-   Other Chess Games (https://www.mersenneforum.org/forumdisplay.php?f=127)
-   -   Engine-assisted game vs Stockfish, move 28 discussion (https://www.mersenneforum.org/showthread.php?t=22100)

MooMoo2 2017-03-02 06:30

Engine-assisted game vs Stockfish, move 28 discussion
 
1 Attachment(s)
The most popular move on the move 27 poll was ...Qxb4. Stockfish played 28. b3, and the FEN is:
2r2r1k/6bp/3p4/3Q1p2/1q2pP2/1P6/4B1PP/3R1R1K b - - 0 28

Game so far: 1.e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 Nc6 6. Ndb5 d6 7. Bf4 e5 8. Bg5 a6 9. Na3 b5 10. Bxf6 gxf6 11. Nd5 f5 12. c3 Bg7 13. Bd3 Be6 14. Nc2 O-O 15. O-O fxe4 16. Bxe4 Rc8 17. a4 f5 18. axb5 axb5 19. Bf3 e4 20. Be2 Kh8 21. Qd2 Be5 22. f4 Bxd5 23. Qxd5 Qb6+ 24. Kh1 Bg7 25. Rad1 b4 26. Nxb4 Nxb4 27. cxb4 Qxb4 28. b3

Depth: 58 ply. Nodes searched: 1,046,398,035,315 (~1046 G)

richs 2017-03-02 17:58

Based on dh1's runs last move, it looks like Rc2. Perhaps dh1 can analyze from the current position?

dh1 2017-03-02 23:12

should have something more in 24hrs; so far:
Houdini at depth 33/72 prefers Rc2 scored at zero.
Komodo at depth 33 has these all four equal scored at zero: Rcd8;Rc2;Qc5;e3.
Shredder at depth 40 has these all five equal scored at zero: Rcd8;Rc2;Qc5;e3;Ra8

WMHalsdorf 2017-03-03 19:37

The Rc2 line is practically a forced line leading to a draw.

dh1 2017-03-04 17:14

none of my Komodos go deep for multi variation >2 which I wanted since several moves are equally rated zero:
Komodo-10.3-64bit:
35 -0.13 28. ... Ra8 29.Bc4 Ra2 30.Rg1 Rf2 31.Qxd6 Qxd6 32.Rxd6 Bc3 33.g4 Rxf4
35 0.00 28. ... Qc5 29.Qxd6 Qxd6 30.Rxd6 Rc2 31.Bh5 Bc3 32.Rd7 Ra8 33.Rfd1 Raa2
35 0.00 28. ... e3 29.Rd3 Qe4 30.Qxe4 fxe4 31.Rxe3 d5 32.Rd1 Rxf4 33.g3 d4
35 0.00 28. ... Bc3 29.g4 fxg4 30.Qxd6 e3 31.Qxb4 Bxb4 32.Bxg4 Rc2 33.Rd4 Rg8
35 0.00 28. ... Rcd8 29.Bc4 Qc5 30.Qxc5 dxc5 31.Rxd8 Rxd8 32.g3 Bd4 33.Be6 Rf8
35 0.00 28. ... Rc2 29.Bc4 Qc5 30.Qxd6 Qxd6 31.Rxd6 Bc3 32.Rfd1 Re8 33.Rd8 Rxd8

Houdini 5:
48/102 -0.45 28. ... Qb6 29.Bc4 Bc3 30.Qxd6 Qxd6 31.Rxd6 Rfe8 32.Rf2 Red8 33.Rxd8+ Rxd8
48/102 -0.35 28. ... Bc3 29.Bc4 Rf6 30.g4 fxg4 31.Qxe4 Rcf8 32.f5 h5 33.h3 Qa5
48/102 -0.23 28. ... Rc2 29.Bc4 Qc5 30.Qe6 Bd4 31.Rc1 Rd2 32.Bd5 Qb5 33.Qxd6 Bg7
48/102 -0.15 28. ... Rcd8 29.Rg1 Qc5 30.Bc4 Bc3 31.g4 Qxd5 32.Rxd5 fxg4 33.Rg3 Bb4
48/102 -0.15 28. ... Qc5 29.Bc4 Rcd8 30.Rg1 Bc3 31.g4 Qxd5 32.Rxd5 fxg4 33.Rg3 Bb4
48/102 -0.08 28. ... e3 29.Qf3 Qe4 30.Rxd6 Rc2 31.h3 Rfc8 32.Rfd1 Bf8 33.Rd8 Qxf3

EngineDeepShredder13UCIx64:
51/83 -0.14 28. ... Bc3 29.Bc4 Qc5 30.Qxd6 Qxd6 31.Rxd6 e3 32.Be2 Rfd8 33.Re6 Bd4
51/83 -0.04 28. ... Ra8 29.Bc4 Ra2 30.h3 Qc3 31.Qxd6 Qc2 32.Rg1 Qf2 33.b4 Rc2
51/83 0.00 28. ... Rcd8 29.Bc4 Qc5 30.Rfe1 Bc3 31.Re2 Qxd5 32.Rxd5 h5 33.Re3 Ra8
51/83 0.00 28. ... Qc5 29.Bc4 Rcd8 30.Rfe1 Bc3 31.Re2 Qxd5 32.Rxd5 h5 33.Re3 Ra8
51/83 0.00 28. ... e3 29.Qxd6 Qxd6 30.Rxd6 Rc2 31.Re1 Rfc8 32.Bd1 Rf2 33.g3 Bf8
51/83 0.00 28. ... Rc2 29.Bc4 Qc5 30.Qxd6 Qxd6 31.Rxd6 Bc3 32.h3 Re8 33.Rfd1 Rf2

The Carnivore 2017-03-05 16:53

[QUOTE=dh1;454238]
Komodo-10.3-64bit:
35 0.00 28. ... Qc5 29.Qxd6 Qxd6 30.Rxd6 Rc2 31.Bh5 Bc3 32.Rd7 Ra8 33.Rfd1 Raa2
35 0.00 28. ... e3 29.Rd3 Qe4 30.Qxe4 fxe4 31.Rxe3 d5 32.Rd1 Rxf4 33.g3 d4
35 0.00 28. ... Rcd8 29.Bc4 Qc5 30.Qxc5 dxc5 31.Rxd8 Rxd8 32.g3 Bd4 33.Be6 Rf8
35 0.00 28. ... Rc2 29.Bc4 Qc5 30.Qxd6 Qxd6 31.Rxd6 Bc3 32.Rfd1 Re8 33.Rd8 Rxd8

EngineDeepShredder13UCIx64:
51/83 0.00 28. ... Rcd8 29.Bc4 Qc5 30.Rfe1 Bc3 31.Re2 Qxd5 32.Rxd5 h5 33.Re3 Ra8
51/83 0.00 28. ... Qc5 29.Bc4 Rcd8 30.Rfe1 Bc3 31.Re2 Qxd5 32.Rxd5 h5 33.Re3 Ra8
51/83 0.00 28. ... e3 29.Qxd6 Qxd6 30.Rxd6 Rc2 31.Re1 Rfc8 32.Bd1 Rf2 33.g3 Bf8
51/83 0.00 28. ... Rc2 29.Bc4 Qc5 30.Qxd6 Qxd6 31.Rxd6 Bc3 32.h3 Re8 33.Rfd1 Rf2[/QUOTE]
[QUOTE=dh1;454238]
Houdini 5:
48/102 -0.23 28. ... Rc2 29.Bc4 Qc5 30.Qe6 Bd4 31.Rc1 Rd2 32.Bd5 Qb5 33.Qxd6 Bg7
48/102 -0.15 28. ... Rcd8 29.Rg1 Qc5 30.Bc4 Bc3 31.g4 Qxd5 32.Rxd5 fxg4 33.Rg3 Bb4
48/102 -0.15 28. ... Qc5 29.Bc4 Rcd8 30.Rg1 Bc3 31.g4 Qxd5 32.Rxd5 fxg4 33.Rg3 Bb4
48/102 -0.08 28. ... e3 29.Qf3 Qe4 30.Rxd6 Rc2 31.h3 Rfc8 32.Rfd1 Bf8 33.Rd8 Qxf3
[/QUOTE]
e3, Qc5, Rcd8, and Rc2 are the only moves rated as 0.00 by both Shredder and Komodo. They're also the 4 best moves suggested by Houdini.

Rc2 is probably not a good idea since it was the worst of the 4 moves according to Houdini. Also, we're playing for a win and not a draw, and WMHalsdorf said it was basically a forced line that leads to a draw.

So now we have e3, Qc5, and Rcd8. e3 is the best move suggested by Houdini, but not by a lot. The other two moves are tied and are both a tiny bit worse than e3.

VBCurtis 2017-03-05 19:11

[QUOTE=The Carnivore;454293]So now we have e3, Qc5, and Rcd8. e3 is the best move suggested by Houdini, but not by a lot. The other two moves are tied and are both a tiny bit worse than e3.[/QUOTE]

I think you're reading the scores backward- they're ranked from best for us; negative scores are in black's favor.

dh1 2017-03-05 23:53

[QUOTE=VBCurtis;454299]I think you're reading the scores backward- they're ranked from best for us; negative scores are in black's favor.[/QUOTE]

The more negative is worse for black as this is a black analysis, and the sort is worse-first.

VBCurtis 2017-03-06 04:16

My apologies! But, why would we want to know the worst moves? Or, why would the worst be situated on top? That means a longer list of moves would necessarily begin with a different move for each list length.

LaurV 2017-03-06 06:02

Some engines use "absolute" scale, in the sense that positive is good for white, negative is good for black. Some other engines use a scale "relative" to who is to move, positive is good for the guy to move, whose move is analyzed, negative is good for the opponent. This surprised me initially too (not a big fan of engine games before these matches), and it was discussed at the starting of this topic. I still don't know how is better, but if asked, I would vote for the absolute way, in spite of the fact that the "relative" way has also its advantages (and disadvantages).

dh1 2017-03-06 09:09

[QUOTE=LaurV;454354] ... Some other engines use a scale "relative" to who is to move, positive is good for the guy to move, whose move is analyzed, negative is good for the opponent. ...[/QUOTE]

this is what I (badly) tried to say. This has to do my my limited understanding of options in the Arena GUI; it shows analysis with the better move on top, but when copied it is highest-ply and better move at the bottom, negative is bad for whichever color is being analysed.

so from my previous, Komodo's best-for-black moves are: the five zero-scores; Houdini best move is e3; and Shredder best moves are: the four zero-scores.

I am open to re-education :-)


All times are UTC. The time now is 03:50.

Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2023, Jelsoft Enterprises Ltd.