Why engines are clueless about this position ?

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

Moderators: hgm, Rebel, chrisw

neelbasant
Posts: 226
Joined: Sun Apr 01, 2012 7:57 pm

Why engines are clueless about this position ?

Post by neelbasant »

[pgn]

1. d4 {248s}{+0.18/30}
{Both last book move} 1... Nf6 {299s}{+0.18/25} 2. c4 {118s}{+0.19/31} 2... c6
{110s}{+0.23/25}{(e6)} 3. Bf4 {274s}{+0.28/31}{(Nc3)} 3... Qb6 {902s}{+0.26/28}
{(d5)} 4. Qd2 {169s}{+0.02/30} 4... Ne4 {217s}{+0.22/27} 5. Qc2 {129s}
{+0.03/32} 5... d5 {279s}{+0.19/28} 6. f3 {163s}{+0.24/32} 6... e5 {218s}
{+0.17/26}{(Nf6)} 7. Bxe5 {112s}{+0.25/34} 7... Qa5+ {69s}{+0.21/26} 8. Nc3
{165s}{+0.19/37} 8... Nxc3 {82s}{+0.16/26} 9. bxc3 {116s}{+0.16/33} 9... dxc4
{274s}{+0.14/27} 10. Bf4 {207s}{+0.20/33}{(e4)} 10... Ba3 {356s}{+0.13/29}
{(Nd7)} 11. e4 {166s}{+0.15/33}{(Rb1)} 11... b5 {163s}{+0.13/28} 12. g3 {364s}
{+0.13/34}{(Be2)} 12... O-O {160s}{+0.13/28} 13. Bh3 {489s}{+0.18/34}{(Ne2)}
13... Be6 {182s}{+0.13/27}{(Bxh3)} 14. Kf2 {284s}{+0.14/33}{(Rb1)} 14... Nd7
{66s}{+0.04/25} 15. Kg2 {77s}{+0.10/32} 15... Rae8 {205s}{-0.01/27}{(Qd8)} 16.
Bxe6 {134s}{+0.03/33} 16... fxe6 {87s}{+0.02/26} 17. Rd1 {168s}[/pgn]
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: Why engines are clueless about this position ?

Post by zullil »

neelbasant wrote:[pgn]

1. d4 {248s}{+0.18/30}
{Both last book move} 1... Nf6 {299s}{+0.18/25} 2. c4 {118s}{+0.19/31} 2... c6
{110s}{+0.23/25}{(e6)} 3. Bf4 {274s}{+0.28/31}{(Nc3)} 3... Qb6 {902s}{+0.26/28}
{(d5)} 4. Qd2 {169s}{+0.02/30} 4... Ne4 {217s}{+0.22/27} 5. Qc2 {129s}
{+0.03/32} 5... d5 {279s}{+0.19/28} 6. f3 {163s}{+0.24/32} 6... e5 {218s}
{+0.17/26}{(Nf6)} 7. Bxe5 {112s}{+0.25/34} 7... Qa5+ {69s}{+0.21/26} 8. Nc3
{165s}{+0.19/37} 8... Nxc3 {82s}{+0.16/26} 9. bxc3 {116s}{+0.16/33} 9... dxc4
{274s}{+0.14/27} 10. Bf4 {207s}{+0.20/33}{(e4)} 10... Ba3 {356s}{+0.13/29}
{(Nd7)} 11. e4 {166s}{+0.15/33}{(Rb1)} 11... b5 {163s}{+0.13/28} 12. g3 {364s}
{+0.13/34}{(Be2)} 12... O-O {160s}{+0.13/28} 13. Bh3 {489s}{+0.18/34}{(Ne2)}
13... Be6 {182s}{+0.13/27}{(Bxh3)} 14. Kf2 {284s}{+0.14/33}{(Rb1)} 14... Nd7
{66s}{+0.04/25} 15. Kg2 {77s}{+0.10/32} 15... Rae8 {205s}{-0.01/27}{(Qd8)} 16.
Bxe6 {134s}{+0.03/33} 16... fxe6 {87s}{+0.02/26} 17. Rd1 {168s}[/pgn]
[D]4rrk1/p2n2pp/2p1p3/qp6/2pPPB2/b1P2PP1/P1Q3KP/3R2NR b - - 1 17
I haven't a clue. What should an engine see here?
User avatar
yurikvelo
Posts: 710
Joined: Sat Dec 06, 2014 1:53 pm

Re: Why engines are clueless about this position ?

Post by yurikvelo »

Stockfish 171114, D=36

1. = (0.00): 17...Nb6 18.Bc7 Qa4 19.Rd2 Qxc2 20.Rxc2 Rf7
2. = (0.00): 17...Rf7 18.Nh3 Nb6 19.Be3 Rd7 20.Rb1 Na4
3. = (0.00): 17...Qd8 18.Ne2 Rf7 19.Rhf1 Qe7 20.h4 Ref8
4. = (0.08): 17...Be7 18.Be3 Ba3 19.Ne2 Qc7 20.Bf4 Qd8
5. = (0.10): 17...a6 18.Nh3 Qd8 19.Nf2 Nb6 20.Rhf1 Na4
6. = (0.12): 17...Qa6 18.Nh3 Nb6 19.Rhf1 Qa5 20.Bc7 Qa4
7. = (0.15): 17...c5 18.d5 exd5 19.exd5 Qa4 20.Qxa4 bxa4
8. = (0.16): 17...Rf6 18.h4 Nb6 19.Nh3 Na4 20.Bd2 Nb2
9. = (0.21): 17...Qb6 18.h4 Qb7 19.h5 h6 20.Nh3 a5
10. = (0.21): 17...Nf6 18.h4 h6 19.Ne2 c5 20.dxc5 Bxc5
User avatar
Thomas Lagershausen
Posts: 328
Joined: Mon Jun 11, 2007 6:59 pm

Re: Why engines are clueless about this position ?

Post by Thomas Lagershausen »

Well, its time to show us the right move in this position.

We know that todays chessprograms are stupid and with little knowledge about strategic chess.

Did you play 17...Rxf4 !?
TL
Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 12:41 pm

Re: Why engines are clueless about this position ?

Post by Lyudmil Tsvetkov »

Very intriguing, non-standard position.

All I am certain is black is better here, although the game might end in a draw.

So it is not 0.0, with some white chances, but rather 0.0, with some black chances.

I think after Rf4 white is better, but a few specific lines might give equality.

I think the strongest move is Nb6 Bc7 Rf7 Bb6 (permitting the knight to land on a4 is even worse for white) ab6

[d]4r1k1/5rpp/1pp1p3/qp6/2pPP3/b1P2PP1/P1Q3KP/3R2NR w - - 0 4

and black is better here.
Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 12:41 pm

Re: Why engines are clueless about this position ?

Post by Lyudmil Tsvetkov »

Lyudmil Tsvetkov wrote:Very intriguing, non-standard position.

All I am certain is black is better here, although the game might end in a draw.

So it is not 0.0, with some white chances, but rather 0.0, with some black chances.

I think after Rf4 white is better, but a few specific lines might give equality.

I think the strongest move is Nb6 Bc7 Rf7 Bb6 (permitting the knight to land on a4 is even worse for white) ab6

[d]4r1k1/5rpp/1pp1p3/qp6/2pPP3/b1P2PP1/P1Q3KP/3R2NR w - - 0 4

and black is better here.
It might be that black is better than I thought.

My best line for the time being after ab6 is: f4 Ra7 Nf3 Bd6 Ra1 Qa6 Ng5 h6 Nf3 Ra8 Rfb1 Qc8

[d]r1q3k1/r5p1/1ppbp2p/1p6/2pPPP2/2P2NP1/P1Q3KP/RR6 w - - 0 12

Is not this won for black?

I must say that I show SF 70-80% of the key moves. It always wants to play something else.
User avatar
yurikvelo
Posts: 710
Joined: Sat Dec 06, 2014 1:53 pm

Re: Why engines are clueless about this position ?

Post by yurikvelo »

Lyudmil Tsvetkov wrote: I must say that I show SF 70-80% of the key moves. It always wants to play something else.
use multiPV
Lyudmil Tsvetkov
Posts: 6052
Joined: Tue Jun 12, 2012 12:41 pm

Re: Why engines are clueless about this position ?

Post by Lyudmil Tsvetkov »

yurikvelo wrote:
Lyudmil Tsvetkov wrote: I must say that I show SF 70-80% of the key moves. It always wants to play something else.
use multiPV
That does not change the best move, and you want to refute or improve precisely on the best move.
User avatar
yurikvelo
Posts: 710
Joined: Sat Dec 06, 2014 1:53 pm

Re: Why engines are clueless about this position ?

Post by yurikvelo »

Lyudmil Tsvetkov wrote:
yurikvelo wrote:
Lyudmil Tsvetkov wrote: I must say that I show SF 70-80% of the key moves. It always wants to play something else.
use multiPV
That does not change the best move, and you want to refute or improve precisely on the best move.
that show how close are other candidate moves, and probability they can become best move with more D.
neelbasant
Posts: 226
Joined: Sun Apr 01, 2012 7:57 pm

Re: Why engines are clueless about this position ?

Post by neelbasant »

Stockfish 171114, D=36

1. = (0.00): 17...Nb6 18.Bc7 Qa4 19.Rd2 Qxc2 20.Rxc2 Rf7
2. = (0.00): 17...Rf7 18.Nh3 Nb6 19.Be3 Rd7 20.Rb1 Na4
3. = (0.00): 17...Qd8 18.Ne2 Rf7 19.Rhf1 Qe7 20.h4 Ref8
4. = (0.08): 17...Be7 18.Be3 Ba3 19.Ne2 Qc7 20.Bf4 Qd8
5. = (0.10): 17...a6 18.Nh3 Qd8 19.Nf2 Nb6 20.Rhf1 Na4
6. = (0.12): 17...Qa6 18.Nh3 Nb6 19.Rhf1 Qa5 20.Bc7 Qa4
7. = (0.15): 17...c5 18.d5 exd5 19.exd5 Qa4 20.Qxa4 bxa4
8. = (0.16): 17...Rf6 18.h4 Nb6 19.Nh3 Na4 20.Bd2 Nb2
9. = (0.21): 17...Qb6 18.h4 Qb7 19.h5 h6 20.Nh3 a5
10. = (0.21): 17...Nf6 18.h4 h6 19.Ne2 c5 20.dxc5 Bxc5


:= 2. .....Rf7 18. Nh3 Nb6..
Here Stockfish did not find Novelty move. 19 .Ng5 Which forces Rxf4 gxf4 h6 leaves black exchange down with no compensation.

Neel
Note.: In game black played 17....Rf7 18.Nh3 Nb6..and after Ng5 it lost.