The Sorcerer's Apprentice

[zullil]

You try to give SF some 10 minutes for a search, and I never use more than 5 min. for thinking on a move.

OK, I think I did not handle that perfectly, just realised that my proposed h4 fails to Ng4 Rg4 Bf5 and the rook can not retreat to h4, as there is already a pawn there - typical human mistake.

White will have to play ef5 Re3 Re3 and that is almost certainly a draw, but that is computer stuff that does not interest us in the least.

For the record, here is the position after your last move, Rg2, along with SF's latest evaluation of that position:
[d]2r1n1k1/ppq1r1pb/3b1p1p/2pPnP1N/P1P1P1P1/1P2QN2/6RP/2B1R1K1 b - - 2 28

Code: Select all

info depth 41 seldepth 65 multipv 1 score cp 116 nodes 98229387004 nps 23007789 hashfull 998 tbhits 0 time 4269397 pv a7a6 h5f4 e5g4 g2g4 h7f5 e4f5 e7e3 c1e3 d6f4 e3f4 c7d7 e1e6 g7g5 h2h4 d7h7 h4g5 h7f5 g4g2 h6g5 f3g5 f5b1 g1h2 f6g5 f4g5 e8g7 g5h6 c8c7 e6g6 g8h7 h6g7 b1g6 g2g6 h7g6 g7f8 b7b6 f8d6 c7d7 d6g3 g6f5 h2h3 f5e4 h3g4 d7f7 g4g5 f7g7 g5h4 e4f5 g3f2 g7d7

Code: Select all

info depth 42 seldepth 65 multipv 1 score cp 125 nodes 132153219771 nps 23189112 hashfull 998 tbhits 0 time 5698934 pv a7a6 h5f4 e5g4 g2g4 h7f5 e4f5 e7e3 c1e3 c7b6 e1b1 g8h7 f3d2 d6f4 g4f4 b6c7 f4e4 e8d6 e3f4 c7d7 f4d6 d7d6 b1f1 d6d7 e4e6 b7b5 a4b5 a6b5 d5d6 b5c4 d2c4 c8d8 e6e3 d7b7 f1d1 h7g8 g1f2 b7d7 d1d5 d7f7 d5c5 f7h5 h2h3 h5d1 b3b4 d1h1 f2g3 h1g1 g3f3 g1f1 f3g3

Not sure you'd even be able to draw from here.

About f5 instead of Qc3, I'll post again soon. If I recall, SF preferred Qe2 instead of Qc3.

[Lyudmil Tsvetkov]

You try to give SF some 10 minutes for a search, and I never use more than 5 min. for thinking on a move.

OK, I think I did not handle that perfectly, just realised that my proposed h4 fails to Ng4 Rg4 Bf5 and the rook can not retreat to h4, as there is already a pawn there - typical human mistake.

White will have to play ef5 Re3 Re3 and that is almost certainly a draw, but that is computer stuff that does not interest us in the least.

For the record, here is the position after your last move, Rg2, along with SF's latest evaluation of that position:
[d]2r1n1k1/ppq1r1pb/3b1p1p/2pPnP1N/P1P1P1P1/1P2QN2/6RP/2B1R1K1 b - - 2 28

Code: Select all

info depth 41 seldepth 65 multipv 1 score cp 116 nodes 98229387004 nps 23007789 hashfull 998 tbhits 0 time 4269397 pv a7a6 h5f4 e5g4 g2g4 h7f5 e4f5 e7e3 c1e3 d6f4 e3f4 c7d7 e1e6 g7g5 h2h4 d7h7 h4g5 h7f5 g4g2 h6g5 f3g5 f5b1 g1h2 f6g5 f4g5 e8g7 g5h6 c8c7 e6g6 g8h7 h6g7 b1g6 g2g6 h7g6 g7f8 b7b6 f8d6 c7d7 d6g3 g6f5 h2h3 f5e4 h3g4 d7f7 g4g5 f7g7 g5h4 e4f5 g3f2 g7d7

Not sure you'd even be able to draw from here.

About f5 instead of Qc3, I'll post again soon.

As said, I will play h4, and not Nf4, which is the weaker move.
I bet h4 draws, but that has no relation to our duscussion of engine weaknesses.

[zullil]

You try to give SF some 10 minutes for a search, and I never use more than 5 min. for thinking on a move.

OK, I think I did not handle that perfectly, just realised that my proposed h4 fails to Ng4 Rg4 Bf5 and the rook can not retreat to h4, as there is already a pawn there - typical human mistake.

White will have to play ef5 Re3 Re3 and that is almost certainly a draw, but that is computer stuff that does not interest us in the least.

For the record, here is the position after your last move, Rg2, along with SF's latest evaluation of that position:
[d]2r1n1k1/ppq1r1pb/3b1p1p/2pPnP1N/P1P1P1P1/1P2QN2/6RP/2B1R1K1 b - - 2 28

Code: Select all

info depth 41 seldepth 65 multipv 1 score cp 116 nodes 98229387004 nps 23007789 hashfull 998 tbhits 0 time 4269397 pv a7a6 h5f4 e5g4 g2g4 h7f5 e4f5 e7e3 c1e3 d6f4 e3f4 c7d7 e1e6 g7g5 h2h4 d7h7 h4g5 h7f5 g4g2 h6g5 f3g5 f5b1 g1h2 f6g5 f4g5 e8g7 g5h6 c8c7 e6g6 g8h7 h6g7 b1g6 g2g6 h7g6 g7f8 b7b6 f8d6 c7d7 d6g3 g6f5 h2h3 f5e4 h3g4 d7f7 g4g5 f7g7 g5h4 e4f5 g3f2 g7d7

Not sure you'd even be able to draw from here.

About f5 instead of Qc3, I'll post again soon.

As said, I will play h4, and not Nf4, which is the weaker move.
I bet h4 draws, but that has no relation to our duscussion of engine weaknesses.

SF believes h4 is weaker than Nf4, but let's leave this line now.

[zullil]

if I had played the stronger move f5 instead of Qc3?

(I had to avoid a possible draw, you know)

Here's a quick analysis after f5.
[d]r2q1rk1/pp1nbpp1/1n4bp/2pP1P2/2P1P3/3Q1N2/PP1N2PP/R1B2RK1 b - - 0 17

Code: Select all

info depth 43 seldepth 66 multipv 1 score cp 53 nodes 18209356302 nps 22008828 hashfull 998 tbhits 0 time 827366 pv g6h7 e4e5 d8c7 f1e1 f8e8 d2b3 e7f8 c1f4 f7f6 d5d6 c7d8 e5e6 d7e5 d3f1 f8d6 f3e5 f6e5 f4e5 d8c7 e5d6 c7d6 g2g4 a8d8 f1f2 b6c4 b3c5 c4d2 a1d1 d2f3 f2f3 d6c5 f3f2 c5c4 h2h3 d8d1 e1d1 c4a2 d1d7 a2b1 g1h2 b1e4 f2a7 e4e5 h2g2 b7b6 a7b6 e5e4 g2g1 h7f5 g4f5 e4f5

Stockfish prefers Qe2 to both Qc3 and f5.

I need to stop now. It seems clear to me that SF is now stronger than any human who has ever played chess---including Bronstein.

[Lyudmil Tsvetkov]

if I had played the stronger move f5 instead of Qc3?

(I had to avoid a possible draw, you know)

Here's a quick analysis after f5.
[d]r2q1rk1/pp1nbpp1/1n4bp/2pP1P2/2P1P3/3Q1N2/PP1N2PP/R1B2RK1 b - - 0 17

Code: Select all

info depth 43 seldepth 66 multipv 1 score cp 53 nodes 18209356302 nps 22008828 hashfull 998 tbhits 0 time 827366 pv g6h7 e4e5 d8c7 f1e1 f8e8 d2b3 e7f8 c1f4 f7f6 d5d6 c7d8 e5e6 d7e5 d3f1 f8d6 f3e5 f6e5 f4e5 d8c7 e5d6 c7d6 g2g4 a8d8 f1f2 b6c4 b3c5 c4d2 a1d1 d2f3 f2f3 d6c5 f3f2 c5c4 h2h3 d8d1 e1d1 c4a2 d1d7 a2b1 g1h2 b1e4 f2a7 e4e5 h2g2 b7b6 a7b6 e5e4 g2g1 h7f5 g4f5 e4f5

Stockfish prefers Qe2 to both Qc3 and f5.

I need to stop now. It seems clear to me that SF is now stronger than any human who has ever played chess---including Bronstein.

And SF is wrong of course.

I saw Bh5 as the only drawing move, that is why I did not play f5, but Qc3 instead.

Bh7 loses.

We might have a line that would end somewhere here:

[d]3rr1k1/ppq3pb/1n1bPp1p/2pP1P2/2P1Q1P1/5N2/PP5P/R1B1R1K1 b - - 0 10

What does SF says here, as black is totally lost: that is the power of the long chains white has, but I simply did not have the stamina left to prove this.

You should not forget that, unlike computers, humans very often get tired, and sometimes even have bad shape periods, like I currently seem to have.
SF is unbothered by any similar concerns.

OK, I do not have a stake in it: all I wanted is to show the power of long chains, so that engine authors could possibly correct their eval.

I will never acknowledge SF is stronger than me, never!

Btw., you must still remember that we have to play a return encounter with me having the white pieces this time. It is for you to tell when you are ready, I am ready any time soon, as obviously, if I can not help with any eval suggestions more now, I can at least beat SF.