why do engines not want to play ...Rxa2 in this position...

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

Moderators: hgm, Rebel, chrisw

pilgrimdan
Posts: 405
Joined: Sat Jul 02, 2011 10:49 pm

why do engines not want to play ...Rxa2 in this position...

Post by pilgrimdan »

[d]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/Pr1N1PPP/R1R3K1 b - - 0 17
Top
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: why do engines not want to play ...Rxa2 in this position

Post by zullil »

pilgrimdan wrote:[d]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/Pr1N1PPP/R1R3K1 b - - 0 17
After Rbxa2:

[D]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/r2N1PPP/R1R3K1 w - - 0 18

Stockfish says

Code: Select all

info depth 41 seldepth 63 score cp 66 nodes 6601576251 nps 9365532 time 704880 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 h2h3 a2a1 g1h2 h7h6 c8c6 f6e8 g2g4 f8d7 h2g2 g8f8 c6c2 d7b6 f3g1 f8e7 g1e2 e7d7 d2f1 g7g6 f1g3 e8c7 e2c3 b6a4 g3e2 a4c3 e2c3 a1a6 c3e2 a6a4 c2b2 d7c6 g2f3 a4a1 e2f4 a1a3 h3h4 c7e6 b2c2 c6d7

info depth 42 seldepth 64 score cp 72 nodes 10666594304 nps 9474696 time 1125798 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 g1f1 h7h6 f1e1 a2a6 c8b8 a6a1 b8b1 a1a6 f3g1 f8d7 g1e2 g8f8 e2c3 d7b6 f2f3 f8e7 e1e2 b6a4 b1b7 e7e6 d2b1 a4c3 b1c3 a6a1 b7b1 a1b1 c3b1 h6h5 b1c3 g7g5 e2d2 e6d7 d2d3 d7e6 g2g3

info depth 43 seldepth 64 score cp 68 nodes 12767421253 nps 9497407 time 1344306 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 g1f1 h7h6 f1e1 a2a6 c8b8 a6a1 d2b1 f6e4 g2g4 g7g6 f3d2 g8g7 e1e2 a1a2 f2f3 e4d2 b1d2 f8e6 h2h4 g7f6 b8b6 f6e7 b6b7 e7e8 b7b8 e6d8 e2d3 e8d7 d2b1 a2f2 b1c3 f2f3 c3d5 f3g3 h4h5 g6h5 g4h5 g3g5 d3c4 d8c6 b8b5
The removal of the Black rook from the back rank and the resulting doubled isolated Black pawns may be the answer to your question.
Top
pilgrimdan
Posts: 405
Joined: Sat Jul 02, 2011 10:49 pm

Re: why do engines not want to play ...Rxa2 in this position

Post by pilgrimdan »

zullil wrote:
pilgrimdan wrote:[d]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/Pr1N1PPP/R1R3K1 b - - 0 17
After Rbxa2:

[D]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/r2N1PPP/R1R3K1 w - - 0 18

Stockfish says

Code: Select all

info depth 41 seldepth 63 score cp 66 nodes 6601576251 nps 9365532 time 704880 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 h2h3 a2a1 g1h2 h7h6 c8c6 f6e8 g2g4 f8d7 h2g2 g8f8 c6c2 d7b6 f3g1 f8e7 g1e2 e7d7 d2f1 g7g6 f1g3 e8c7 e2c3 b6a4 g3e2 a4c3 e2c3 a1a6 c3e2 a6a4 c2b2 d7c6 g2f3 a4a1 e2f4 a1a3 h3h4 c7e6 b2c2 c6d7

info depth 42 seldepth 64 score cp 72 nodes 10666594304 nps 9474696 time 1125798 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 g1f1 h7h6 f1e1 a2a6 c8b8 a6a1 b8b1 a1a6 f3g1 f8d7 g1e2 g8f8 e2c3 d7b6 f2f3 f8e7 e1e2 b6a4 b1b7 e7e6 d2b1 a4c3 b1c3 a6a1 b7b1 a1b1 c3b1 h6h5 b1c3 g7g5 e2d2 e6d7 d2d3 d7e6 g2g3

info depth 43 seldepth 64 score cp 68 nodes 12767421253 nps 9497407 time 1344306 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 g1f1 h7h6 f1e1 a2a6 c8b8 a6a1 d2b1 f6e4 g2g4 g7g6 f3d2 g8g7 e1e2 a1a2 f2f3 e4d2 b1d2 f8e6 h2h4 g7f6 b8b6 f6e7 b6b7 e7e8 b7b8 e6d8 e2d3 e8d7 d2b1 a2f2 b1c3 f2f3 c3d5 f3g3 h4h5 g6h5 g4h5 g3g5 d3c4 d8c6 b8b5
The removal of the Black rook from the back rank and the resulting doubled isolated Black pawns may be the answer to your question.
using the other rook... there doesn't seem to be any problem...

[d]6k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/rr1N1PPP/R1R3K1 w - - 0 18
Top
bob
Posts: 20943
Joined: Mon Feb 27, 2006 7:30 pm
Location: Birmingham, AL

Re: why do engines not want to play ...Rxa2 in this position

Post by bob »

pilgrimdan wrote:
zullil wrote:
pilgrimdan wrote:[d]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/Pr1N1PPP/R1R3K1 b - - 0 17
After Rbxa2:

[D]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/r2N1PPP/R1R3K1 w - - 0 18

Stockfish says

Code: Select all

info depth 41 seldepth 63 score cp 66 nodes 6601576251 nps 9365532 time 704880 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 h2h3 a2a1 g1h2 h7h6 c8c6 f6e8 g2g4 f8d7 h2g2 g8f8 c6c2 d7b6 f3g1 f8e7 g1e2 e7d7 d2f1 g7g6 f1g3 e8c7 e2c3 b6a4 g3e2 a4c3 e2c3 a1a6 c3e2 a6a4 c2b2 d7c6 g2f3 a4a1 e2f4 a1a3 h3h4 c7e6 b2c2 c6d7

info depth 42 seldepth 64 score cp 72 nodes 10666594304 nps 9474696 time 1125798 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 g1f1 h7h6 f1e1 a2a6 c8b8 a6a1 b8b1 a1a6 f3g1 f8d7 g1e2 g8f8 e2c3 d7b6 f2f3 f8e7 e1e2 b6a4 b1b7 e7e6 d2b1 a4c3 b1c3 a6a1 b7b1 a1b1 c3b1 h6h5 b1c3 g7g5 e2d2 e6d7 d2d3 d7e6 g2g3

info depth 43 seldepth 64 score cp 68 nodes 12767421253 nps 9497407 time 1344306 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 g1f1 h7h6 f1e1 a2a6 c8b8 a6a1 d2b1 f6e4 g2g4 g7g6 f3d2 g8g7 e1e2 a1a2 f2f3 e4d2 b1d2 f8e6 h2h4 g7f6 b8b6 f6e7 b6b7 e7e8 b7b8 e6d8 e2d3 e8d7 d2b1 a2f2 b1c3 f2f3 c3d5 f3g3 h4h5 g6h5 g4h5 g3g5 d3c4 d8c6 b8b5
The removal of the Black rook from the back rank and the resulting doubled isolated Black pawns may be the answer to your question.
using the other rook... there doesn't seem to be any problem...

[d]6k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/rr1N1PPP/R1R3K1 w - - 0 18
Crafty actually plays that rook capture, it still thinks white is a half-pawn better, however...
Top
pilgrimdan
Posts: 405
Joined: Sat Jul 02, 2011 10:49 pm

Re: why do engines not want to play ...Rxa2 in this position

Post by pilgrimdan »

bob wrote:
pilgrimdan wrote:
zullil wrote:
pilgrimdan wrote:[d]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/Pr1N1PPP/R1R3K1 b - - 0 17
After Rbxa2:

[D]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/r2N1PPP/R1R3K1 w - - 0 18

Stockfish says

Code: Select all

info depth 41 seldepth 63 score cp 66 nodes 6601576251 nps 9365532 time 704880 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 h2h3 a2a1 g1h2 h7h6 c8c6 f6e8 g2g4 f8d7 h2g2 g8f8 c6c2 d7b6 f3g1 f8e7 g1e2 e7d7 d2f1 g7g6 f1g3 e8c7 e2c3 b6a4 g3e2 a4c3 e2c3 a1a6 c3e2 a6a4 c2b2 d7c6 g2f3 a4a1 e2f4 a1a3 h3h4 c7e6 b2c2 c6d7

info depth 42 seldepth 64 score cp 72 nodes 10666594304 nps 9474696 time 1125798 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 g1f1 h7h6 f1e1 a2a6 c8b8 a6a1 b8b1 a1a6 f3g1 f8d7 g1e2 g8f8 e2c3 d7b6 f2f3 f8e7 e1e2 b6a4 b1b7 e7e6 d2b1 a4c3 b1c3 a6a1 b7b1 a1b1 c3b1 h6h5 b1c3 g7g5 e2d2 e6d7 d2d3 d7e6 g2g3

info depth 43 seldepth 64 score cp 68 nodes 12767421253 nps 9497407 time 1344306 multipv 1 pv a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 g1f1 h7h6 f1e1 a2a6 c8b8 a6a1 d2b1 f6e4 g2g4 g7g6 f3d2 g8g7 e1e2 a1a2 f2f3 e4d2 b1d2 f8e6 h2h4 g7f6 b8b6 f6e7 b6b7 e7e8 b7b8 e6d8 e2d3 e8d7 d2b1 a2f2 b1c3 f2f3 c3d5 f3g3 h4h5 g6h5 g4h5 g3g5 d3c4 d8c6 b8b5
The removal of the Black rook from the back rank and the resulting doubled isolated Black pawns may be the answer to your question.
using the other rook... there doesn't seem to be any problem...

[d]6k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/rr1N1PPP/R1R3K1 w - - 0 18
Crafty actually plays that rook capture, it still thinks white is a half-pawn better, however...
if the goal of black (in this position) is to obtain a draw...

wouldn't the most direct, efficient way of doing this be by taking the pawn...

seems like a human playing black would take the pawn without hesitation...
Top
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: why do engines not want to play ...Rxa2 in this position

Post by zullil »

pilgrimdan wrote: using the other rook... there doesn't seem to be any problem...

[d]6k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/rr1N1PPP/R1R3K1 w - - 0 18
Did you glance at the line from Stockfish in my previous post?

The choice of rooks is irrelevant, since Stockfish will exchange rooks immediately in either case, leading to this:

[D]6k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/r2N1PPP/2R3K1 w - - 0 19

Code: Select all

info depth 35 seldepth 52 score cp 82 nodes 1111217853 nps 9123299 time 121800 multipv 1 pv c4d5 e6d5 c1c8 d7f8 g1f1 a2b2 f1e1 g7g6 c8c6 f6e8 e1e2 f8e6 f3e1 b2b7 e1d3 g8f8 c6a6 f8e7 a6a5 e8c7 g2g4 c7b5 h2h4 b5c3 e2f3 h7h6 a5a6 c3b5 a6a8 b5c3 a8a5 c3b5 a5a1 b5c3 a1a5
Stockfish sees this position as favorable enough for White that it avoids Rxa2 in the original position.

Here's Stockfish's analysis of the original position (8 threads, 4 GB hash, depth 35, MultiPV = 3):

Code: Select all

info depth 35 seldepth 55 score cp -26 nodes 3487091059 nps 8679602 time 401757 multipv 1 pv f6e4 d2e4 d5e4 f3g5 d7f6 a2a4 h7h6 g5h3 g8f8 a4a5 f8e7 a1a3 g7g5 a5a6 a8a7 f2f4 e4f3 g2f3 f6e8 f3f4 f7f6 h3f2 b2b6 c1a1 e8c7 f2d3 b6a6 a3a6 a7a6 a1a6 c7a6 g1f2 a6c7 f4g5 h6g5 f2f3 f6f5 e3e4 e7d7 c4c5 g5g4 f3f4

info depth 35 seldepth 55 score cp -63 nodes 3487091059 nps 8679602 time 401757 multipv 2 pv g7g5 h2h3 b2a2 a1a2 a8a2 c4d5 e6d5 d2b1 f6e4 b1c3 a2a5 c3e4 d5e4 f3d2 d6d5 g2g4 d7f6 d2f1 h7h5 c1c6 g8g7 f1g3 h5g4 g3f5 g7g6 h3g4 a5a1 g1g2 a1a2 g2g3 a2a1 f2f3 e4f3 g3f3 a1f1

info depth 35 seldepth 55 score cp -75 nodes 3487091059 nps 8679602 time 401757 multipv 3 pv b2a2 a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 g1f1 g7g5 f1e1 h7h6 c8c6 g8g7 h2h3 f6e8 d2b1 f8e6 b1c3 a2c2 f3d2 e6c7 e1d1 c2b2 d1c1 b2b4 c1c2 g7f6 c3e2 f6e7 c2d3 h6h5 c6c1 c7b5 c1a1 b4b2 a1a5 h5h4 e2c3 e8c7 c3b5
info nodes 3487091059 time 401757
bestmove f6e4 ponder d2e4
Stockfish simply "believes" Ne4 is clearly better than either Rook capturing the pawn on a2.
Top
Vinvin
Posts: 5228
Joined: Thu Mar 09, 2006 9:40 am
Full name: Vincent Lejeune

Re: why do engines not want to play ...Rxa2 in this position

Post by Vinvin »

pilgrimdan wrote:[d]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/Pr1N1PPP/R1R3K1 b - - 0 17
SF likes Ne4 after 30 min on an Intel 3.1 ghz, 2 cores :

FEN: r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/Pr1N1PPP/R1R3K1 b - - 0 1

Note that "-" means white advantage in Arena when black to play.

FEN: r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/Pr1N1PPP/R1R3K1 b - - 0 1

Stockfish_14011911_32bit:
...
28/52 01:38 210.293.318 2.129.423 -0,67 g5 h3 h5 cxd5 exd5 Td1 g4 hxg4 hxg4 Ch4 Rh7 a4 Cb6 Cf5 Cxa4 Cxd6 Cc3 Txa8 Cxd1 Cf1 Rg6 f3 gxf3 gxf3 Tb6 Ta1 Cxe3 Cxe3 Txd6 Ta5 Tb6 Cxd5
28/52 01:38 210.293.318 2.129.423 -0,62 Taxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 Ce4 Cxe4 dxe4 Cg5 Ta1+ Re2 d5 Ch3 Ta5 f3 h6 fxe4 dxe4 Cf4 g5 Ch5 Ta2+ Rf1 Ta1+ Rf2 Ta2+ Rg3 Te2 Cf6+ Rg7 Ce8+ Rg8 Tc3 Cd7 Ta3 h5 h3 h4+ Rg4 Txg2+ Rf5
28/52 01:38 210.293.318 2.129.423 -0,62 Tbxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 Ce4 Cxe4 dxe4 Cg5 Ta1+ Re2 d5 Ch3 Ta5 f3 h6 fxe4 dxe4 Cf4 g5 Ch5 Ta2+ Rf1 Ta1+ Rf2 Ta2+ Rg3 Te2 Cf6+ Rg7 Ce8+ Rg8 Tc3 Cd7 Ta3 h5 h3 h4+ Rg4 Txg2+ Rf5
28/52 01:38 210.293.318 2.129.423 -0,59 Ce4 Cxe4 dxe4 Cg5 Cf6 a4 h6 Ch3 Rf8 a5 Re7 Ta3 g5 a6 Ta7 f4 exf3 gxf3 Tb6 Tca1 Cd7 f4 gxf4 Cxf4 Tc6 T3a2 Txc4 Rf2 Cf6 Tb1 Tc6 Tb7+
---------------------------------------------------------------------------
29/52 02:02 260.729.807 2.136.096 -0,69 Taxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 Ce4 Cxe4 dxe4 Cg5 Ta1+ Re2 Ta2+ Re1 d5 Ch3 Ta5 g3 g6 Rf1 Rg7 Cf4 Ce6 Cxe6+ fxe6 Tc7+ Rg8 Te7 Ta6 g4 h5 gxh5 gxh5 Rg2 Rf8 Th7 Ta2 Rg3
29/52 02:02 260.729.807 2.136.096 -0,69 Tbxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 Ce4 Cxe4 dxe4 Cg5 Ta1+ Re2 Ta2+ Re1 d5 Ch3 Ta5 g3 g6 Rf1 Rg7 Cf4 Ce6 Cxe6+ fxe6 Tc7+ Rg8 Te7 Ta6 g4 h5 gxh5 gxh5 Rg2 Rf8 Th7 Ta2 Rg3
29/52 02:02 260.729.807 2.136.096 -0,67 g5 h3 h5 cxd5 exd5 Td1 g4 hxg4 hxg4 Ch4 Rh7 a4 Cb6 Cf5 Cxa4 Cxd6 Cc3 Txa8 Cxd1 Cf1 Rg6 f3 gxf3 gxf3 Tb6 Ta1 Cxe3 Cxe3 Txd6 Ta5 Tb6 Cxd5
29/52 02:02 260.729.807 2.136.096 -0,39 Ce4 Cxe4 dxe4 Cg5 Cf6 a4 h6 Ch3 Rf8 a5 g5 Ta3 e5 f4 exf3 gxf3 e4 fxe4 Cxe4 a6 Re7 c5 Ta7 c6 Rd8 Tf1 Tc2 d5 Td2 Cf2 Cxf2
---------------------------------------------------------------------------
30/52 02:45 353.017.173 2.135.886 -0,69 Tbxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 Ce4 Cxe4 dxe4 Cg5 Ta1+ Re2 Ta2+ Re1 d5 Ch3 Ta5 g3 g6 Rf1 Rg7 Cf4 Ce6 Cxe6+ fxe6 Tc7+ Rg8 Te7 Ta6 g4 h5 gxh5 gxh5 Rg2 Rf8 Th7 Ta2 Rg3
30/52 02:45 353.017.173 2.135.886 -0,67 Taxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 g6 Re1 Ta6 Cb1 Rg7 Cfd2 Ce6 Cc3 Ta1+ Re2 Tg1 g3 Tg2 h3 Cg5 h4 Ch3 Cd1 Cg4 Rf3 Tg1 Cc3 Ch2+ Re2 Cg4 Cxd5
30/52 02:45 353.017.173 2.135.886 -0,67 g5 h3 h5 cxd5 exd5 Td1 g4 hxg4 hxg4 Ch4 Rh7 a4 Cb6 Cf5 Cxa4 Cxd6 Cc3 Txa8 Cxd1 Cf1 Rg6 f3 gxf3 gxf3 Tb6 Ta1 Cxe3 Cxe3 Txd6 Ta5 Tb6
30/52 02:45 353.017.173 2.135.886 -0,13 Ce4 Cxe4 dxe4 Cg5 Cf6 a4 h6 Ch3 Rf8 a5 g5 Ta4 Re7 a6 Ta7 Tca1 Rd7 T4a2 Txa2 Txa2 Rc6 f3 exf3 gxf3 Rb6 Cf2 Txa6 Txa6+ Rxa6 e4 Rb6 Cg4 Cxg4 fxg4 e5
---------------------------------------------------------------------------
31/53 05:46 742.954.134 2.146.503 -0,67 Taxa2 Txa2 Txa2 c5 dxc5
31/53 05:46 742.954.134 2.146.503 -0,65 Tbxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 g6 Re1 Ta6 Cb1 Ce4 Cfd2 f5 f3 Cf6 Cc3 Rf7 Re2 g5 Tb8 C8d7 Tb5 Re6 h3 f4 h4 gxh4 exf4 Rf5
31/53 05:46 742.954.134 2.146.503 -0,51 g5 cxd5 exd5 Cf1 Ce4 a4 Cb6 Tcb1 Txb1 Txb1 Cxa4 Ta1 Cb6 Txa8+ Cxa8 h4 g4 C3d2 Cxd2 Cxd2 Cc7 Cb1 Cb5 Rf1 Rg7 Re2 f5 Rd3 Rf6
31/53 05:46 742.954.134 2.146.503 -0,22 Ce4 Cxe4 dxe4 Cg5 Cf6 a4 h6 Ch3 Rf8 a5 g5 Ta3 Re7 a6 Ta7 f3 exf3 gxf3 Tb6 Tca1 Tb4 c5 dxc5 dxc5 Tc4 Cf2 Txc5 Cd3 Tc6 Ce5 Tc5 Ta5 Txa5
---------------------------------------------------------------------------
32/53 10:41 1.329.430.447 2.072.701 -0,84 Tbxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 h6 Tc6 Ta1+ Re2 Ta2 Ce1 Ce4 Tc2 f5 Txa2 Cc3+ Rd3 Cxa2 Cf1 Rf7 Cg3 g6 Cc2 Cc1+ Rd2 Ca2 Ce2 Ce6 Rd3
32/53 10:41 1.329.430.447 2.072.701 -0,81 Rf8 cxd5 exd5 a3 Cb6 h3 Ta7 Rf1 Re7 Re1 Re6 g4 h6 Cg1 Ce4 Cxe4 dxe4 Ce2 g6 Cc3 f5 gxf5+ gxf5 Tcb1 Cc4 a4 Tc2 Ce2 Cb2 Cf4+
32/53 10:41 1.329.430.447 2.072.701 -0,61 g5 cxd5 exd5 Td1 h6 a4 Ta5 h4 g4 Ce1 Ce4 Cxe4 dxe4 Tdc1 Rf8 g3 Cb6 Cg2 Cd5 Tc8+ Re7 Th8 Tb4 Txh6 Tbxa4 Txa4 Txa4 Th5 Ta1+ Rh2 Cf6 Tb5 Ta2 Tb7+ Rf8
32/53 10:41 1.329.430.447 2.072.701 -0,20 Ce4 Cxe4 dxe4 Cg5 Cf6 a4 h6 Ch3 g5 a5 Rf8 Ta3 Re7 a6 Ta7 f3 exf3 gxf3 Tb6 Tca1 Tb4 c5 dxc5 dxc5 Tc4 Cf2 Txc5 Cd3 Tc2 Cb4 Tc4 Tb1 Cd7 Rg2 g4 Ta5 f5
---------------------------------------------------------------------------
33/54 16:48 2.135.458.916 2.117.806 -0,78 Tbxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 g5 Re1 h6 Tc6 Ta1+ Re2 Ta2 Ce1 Ce4 Tc2
33/54 16:48 2.135.458.916 2.117.806 -0,77 Taxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 g5 Re1 h6 Tc6 Ta1+ Re2 Ta2 Ce1 Ce4 Tc2 Rg7 Txa2 Cc3+ Rd3 Cxa2 Rc2 Cb4+ Rb3 Cc6 Cb1 Rf6 Cc3 Re6 Cb5 Rd7 Cd3 f5 h3 Ce6 Cc3 Cc7 Cb4 Ce7
33/54 16:48 2.135.458.916 2.117.806 -0,65 g5 cxd5 exd5 h3 h5 Cb1 Ce4 Cc3 Cdf6 a4 Cxc3 Txc3 Ce4 Tca3 g4 Ce1 Cxf2 a5 gxh3 a6 Rf8 a7 hxg2 Cxg2 Re7 Cf4 Ce4 T1a2 Txa2
33/54 16:48 2.135.458.916 2.117.806 -0,12 Ce4 Cxe4 dxe4 Cg5 Cf6 a4 h6 Ch3 g5 a5 Rf8 Ta3 Re7 a6 Ta7 Ta5 Rd7 g3 Rc6 Rg2 Rb6 Tca1 Cg4 Rg1 Cf6 T5a4 Rc6 Td1 Cg4 Tc1 Cf6 Rg2 Cg4 Ta3 Cf6
---------------------------------------------------------------------------
34/54 21:22 2.734.570.906 2.132.400 -0,84 Tbxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 h6 Tc6 Ta1+ Re2 Ta2 Ce1 Ce4 Tc2 f5 Txa2 Cc3+ Rd3 Cxa2 Cf1 Rf7 Cg3 g6 Cc2 Ce6 Ce2 Rf6
34/54 21:22 2.734.570.906 2.132.400 -0,80 Taxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 g5 Re1 h6 Tc6 Ta1+ Re2 Ta2 Ce1 Ce4 Cc2 Cg6 Cxe4 dxe4 Rd2 Ch4 Txd6 Rg7 d5 Cxg2 Rc3 Ch4
34/54 21:22 2.734.570.906 2.132.400 -0,69 g5 h3 Tbxa2 Txa2 Txa2 cxd5 exd5 Tc6 g4 hxg4 Cxg4 Cf1 Cdf6 C3d2 Ce4 Cxe4 dxe4 f3 exf3 gxf3 Cf2 Rg2 d5 Rg3 Te2 Tc5 Cd1 Txd5 Cxe3 Cxe3 Txe3 Rf4 Tb3 Tg5+ Rf8 Tc5 Rg7 Tc6
34/54 21:22 2.734.570.906 2.132.400 -0,17 Ce4 Cxe4 dxe4 Cg5 Cf6 a4 h6 Ch3 g5 a5 Rf8 Ta3 Re7 a6 Ta7 f4 exf3 gxf3 Tb6 Tca1 Tb4 c5 dxc5 dxc5 Tc4 Cf2 Txc5 Cd3 Tc6 Ce5 Tb6 Rf2 Cd5 e4 Tb2+ Rg1 Cb4 Ta4 Rf6
---------------------------------------------------------------------------
35/56 26:21 3.388.817.203 2.143.368 -0,79 Taxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 g5 Re1 h6 Cb1 Rg7 Cfd2 Cg6 Cc3 Ta1+ Re2 Ch4 g3 Cf5 Cdb1 g4 Tb8 h5 Tb5 Ce7 Tb7 Cf5 Rd3 Ta5 Tb6 Ta1
35/56 26:21 3.388.817.203 2.143.368 -0,76 Tbxa2 Txa2 Txa2 cxd5 exd5 Tc8+ Cf8 Rf1 g5 Re1 h6 Cb1 Rg7 Cfd2 Cg6 Cc3 Ta1+ Re2 Ch4 g3 Cf5 Cdb1 g4 Tb8 Rg6 Tb6 Ta5 Tb5 Ta1 Cd2 Ce7
35/56 26:21 3.388.817.203 2.143.368 -0,49 g5 h3 Tbxa2 Txa2 Txa2 c5 dxc5 dxc5 Rg7 c6 Cb6 Cb3 h6 Cbd4 Ce8 Ce5 Cc8 Cd7 h5 Tb1 Ccd6 Tb7 Ta1+ Rh2 Ta2 c7 Ta8 Cb6 Cxb7 Cxa8 Cbd6
35/56 26:21 3.388.817.203 2.143.368 -0,13 Ce4 Cxe4 dxe4 Cg5 Cf6 a4 h6 Ch3 g5 a5 Rf8 Ta3 Re7 a6 Ta7 f3 exf3 gxf3 Tb6 Tca1 Tb4 c5 dxc5 dxc5 Tc4 Cf2 Txc5 Cd3 Tc6 Ce5 Tb6 Rg2 Cd5 e4 Cf4+ Rf2 Tb2+ Re3
Top
zullil
Posts: 6442
Joined: Tue Jan 09, 2007 12:31 am
Location: PA USA
Full name: Louis Zulli

Re: why do engines not want to play ...Rxa2 in this position

Post by zullil »

pilgrimdan wrote:[d]r5k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/Pr1N1PPP/R1R3K1 b - - 0 17
Perhaps this could be a good test position. I just built the latest Stockfish from github, and it certainly feels the drawing move for Black here is Ne4. Looks pretty good to me too, though I'm not a strong player:

Code: Select all

info depth 46 seldepth 58 score cp 0 nodes 11273777945 nps 9320445 time 1209575 multipv 1 pv f6e4 d2e4 d5e4 f3g5 d7f6 f2f3 e4f3 g5f3 a8a2 a1a2 b2a2 c4c5 d6c5 d4c5 g8f8 c5c6 a2a8 f3d4 f8e7 g1f2 f6d5 e3e4 d5c7 c1d1 a8a6 f2e3 a6a3 e3f4 e7f6 e4e5 f6e7 d4f5 e6f5 d1d7 e7e6 d7d6 e6e7 d6d7

info depth 47 seldepth 58 score cp 0 nodes 14069745848 nps 9424149 time 1492946 multipv 1 pv f6e4 d2e4 d5e4 f3g5 d7f6 f2f3 e4f3 g5f3 a8a2 a1a2 b2a2 c4c5 d6c5 d4c5 g8f8 c5c6 a2a8 f3d4 f8e7 g1f2 f6d5 e3e4 d5c7 c1d1 a8a6 f2e3 a6a3 e3f4 e7f6 e4e5 f6e7 d4f5 e6f5 d1d7 e7e6 d7d6 e6e7 d6d7

info depth 48 seldepth 58 score cp 0 nodes 17339698104 nps 9492107 time 1826749 multipv 1 pv f6e4 d2e4 d5e4 f3g5 d7f6 f2f3 e4f3 g5f3 a8a2 a1a2 b2a2 c4c5 d6c5 d4c5 g8f8 c5c6 a2a8 f3d4 f8e7 g1f2 f6d5 e3e4 d5c7 c1d1 a8a6 f2e3 a6a3 e3f4 e7f6 e4e5 f6e7 d4f5 e6f5 d1d7 e7e6 d7d6 e6e7 d6d7
Note that Black gets the a2 pawn anyway, but with a better position to boot.
Top
pilgrimdan
Posts: 405
Joined: Sat Jul 02, 2011 10:49 pm

Re: why do engines not want to play ...Rxa2 in this position

Post by pilgrimdan »

zullil wrote:
pilgrimdan wrote: using the other rook... there doesn't seem to be any problem...

[d]6k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/rr1N1PPP/R1R3K1 w - - 0 18
Did you glance at the line from Stockfish in my previous post?

The choice of rooks is irrelevant, since Stockfish will exchange rooks immediately in either case, leading to this:

[D]6k1/3n1ppp/3ppn2/3p4/2PP4/4PN2/r2N1PPP/2R3K1 w - - 0 19

Code: Select all

info depth 35 seldepth 52 score cp 82 nodes 1111217853 nps 9123299 time 121800 multipv 1 pv c4d5 e6d5 c1c8 d7f8 g1f1 a2b2 f1e1 g7g6 c8c6 f6e8 e1e2 f8e6 f3e1 b2b7 e1d3 g8f8 c6a6 f8e7 a6a5 e8c7 g2g4 c7b5 h2h4 b5c3 e2f3 h7h6 a5a6 c3b5 a6a8 b5c3 a8a5 c3b5 a5a1 b5c3 a1a5
Stockfish sees this position as favorable enough for White that it avoids Rxa2 in the original position.

Here's Stockfish's analysis of the original position (8 threads, 4 GB hash, depth 35, MultiPV = 3):

Code: Select all

info depth 35 seldepth 55 score cp -26 nodes 3487091059 nps 8679602 time 401757 multipv 1 pv f6e4 d2e4 d5e4 f3g5 d7f6 a2a4 h7h6 g5h3 g8f8 a4a5 f8e7 a1a3 g7g5 a5a6 a8a7 f2f4 e4f3 g2f3 f6e8 f3f4 f7f6 h3f2 b2b6 c1a1 e8c7 f2d3 b6a6 a3a6 a7a6 a1a6 c7a6 g1f2 a6c7 f4g5 h6g5 f2f3 f6f5 e3e4 e7d7 c4c5 g5g4 f3f4

info depth 35 seldepth 55 score cp -63 nodes 3487091059 nps 8679602 time 401757 multipv 2 pv g7g5 h2h3 b2a2 a1a2 a8a2 c4d5 e6d5 d2b1 f6e4 b1c3 a2a5 c3e4 d5e4 f3d2 d6d5 g2g4 d7f6 d2f1 h7h5 c1c6 g8g7 f1g3 h5g4 g3f5 g7g6 h3g4 a5a1 g1g2 a1a2 g2g3 a2a1 f2f3 e4f3 g3f3 a1f1

info depth 35 seldepth 55 score cp -75 nodes 3487091059 nps 8679602 time 401757 multipv 3 pv b2a2 a1a2 a8a2 c4d5 e6d5 c1c8 d7f8 g1f1 g7g5 f1e1 h7h6 c8c6 g8g7 h2h3 f6e8 d2b1 f8e6 b1c3 a2c2 f3d2 e6c7 e1d1 c2b2 d1c1 b2b4 c1c2 g7f6 c3e2 f6e7 c2d3 h6h5 c6c1 c7b5 c1a1 b4b2 a1a5 h5h4 e2c3 e8c7 c3b5
info nodes 3487091059 time 401757
bestmove f6e4 ponder d2e4
Stockfish simply "believes" Ne4 is clearly better than either Rook capturing the pawn on a2.
okay.. thanks... I was looking at it from the standpoint of Black looking to get a draw in a straight forward manner...
Top

Return to “Computer Chess Club: General Topics”