Human versus Machine

Discussion of computer chess matches and engine tournaments.

Moderators: hgm, Harvey Williamson, bob

fastgm
Posts: 334
Joined: Mon Aug 19, 2013 4:57 pm
Contact:

Re: Stockfish and Fischer

Post by fastgm » Wed Jan 17, 2018 10:55 pm

Lyudmil Tsvetkov wrote:If you/SF is able to take back moves, I should be able to do the same, I guess.
In your line, e4 instead of Rg1 seems to be much better:

[d]1rr2bk1/3q3p/4pRp1/p2pP1P1/2pPP3/P1P3RQ/3B3P/3bN2K b - - 0 5

Maybe, black will be able to draw, I don't know.
[d]r1q2rk1/pb2bp1p/4p1p1/1p1pP1Pn/2pP4/P1P1P2Q/1PBN3P/R1B2RK1 w - - 0 17
To remind the whole move order:
17. Rf6 Qd8 18. Nf3 a5 19. Bd2 b4 20. Kh1 Bc6 21. Rg1 bxa3 22. bxa3 Rb8 23. Ne1 Qd7 24. Bd1 Ng7 25. Qh6 Ba4 26. Bg4 Rfc8 27. Rg3 Nh5 28. Bxh5 Bf8 29. Bxg6 fxg6 30. Qh3 Bd1 31. e4

Stockfish is optimistic to hold the position or even to win it after 31.e4:

[d]1rr2bk1/3q3p/4pRp1/p2pP1P1/2pPP3/P1P3RQ/3B3P/3bN2K b - - 0 31

Code: Select all

 48/85	1:03:24	139.396.477k	36.644k	+2,22	31. ... dxe4 32.Rg2 Bxa3 33.Rgf2 Bh5 34.Kg2 Rb1 35.Qe3 Bg4 36.Qxe4 Qb7 37.Qxb7 Rxb7 38.Nc2 Be7 39.R6f4 Bf5 40.Ne3 Bxg5 41.Nxf5 gxf5 42.R4f3 Bxd2 43.Rxd2 a4 44.d5 a3 45.Rff2 exd5 46.Rxd5 Ra8 47.Ra2 Rb3 48.e6 Rxc3 49.e7 Re8 50.Ra5 Rxe7 51.R5xa3 Rxa3 52.Rxa3 Kf7 53.Kf3 Kf6 54.Ra2 c3 55.Rc2 Rc7 56.Kf4 h6 57.h3 Ke6 58.Ke3 Ke5 59.Kd3 f4 60.Rxc3
Lyudmil Tsvetkov wrote:I don't know what you want to claim, that Rf6 does not win?
Certainly it does, I am just too tired now and extremely busy to do careful analysis of alternative lines.
So far I have not claimed anything, have I?
It's an very interesting and highly complex position. Rf6 is the best move and gives white advantage, but whether this advantage is enough to win is still not quite clear. Right now I would say white has good winning chances.
Lyudmil Tsvetkov wrote:Btw., after Rf6 Qd8 Nf1 Nf6 gf6 Bf6 ef6 Qf6, did you check with longer search all the lines:

[d]r4rk1/pb3p1p/4pqp1/1p1p4/2pP4/P1P1P2Q/1PB4P/R1B2NK1 w - - 0 5

I guess this is won for white, but you will need 30-40 moves until the score changes.
I myself tried this with SF and it starts winning after 30-40 moves.
From my experience drawn from different databases and top engine games, 2 minor pieces vs rook + 2 pawns, when the 2 minors also have the bishop pair is won for the minor side, in maybe at least 90% of cases, so, I suggest you

check more carefully/maybe play some games to get convinced.
I'll take again a closer look at this position.
Lyudmil Tsvetkov wrote:Alternatively, white has also Bd1 instead of Nf1:

[d]r2q1rk1/pb2bp1p/4pRp1/1p1pP1Pn/2pP4/P1P1P2Q/1P1N3P/R1BB2K1 b - - 0 2

This should also be investigated closely.

I am fully certain Rf6 wins, and maybe other moves too.
I just don't have even couple of minutes time now to give just this position.

It is a pity all the most beautiful lines are not evident from what SF chooses.
There are really some very intriguing variations here, if SF accepts the sacrifice.
Pity I don't have more time.
As I said before, it's a very complex position.
The white attack looks very dangerous. But black also has defense resources and counterplay on the queenside.

The whole thread has become quite confusing and it is becoming more and more difficult to follow the postings and analyses.
Maybe I have the time for a summary of our analyses.

Lyudmil Tsvetkov
Posts: 6033
Joined: Tue Jun 12, 2012 10:41 am

Re: Stockfish and Fischer

Post by Lyudmil Tsvetkov » Sat Jan 20, 2018 7:21 pm

fastgm wrote:
Lyudmil Tsvetkov wrote:If you/SF is able to take back moves, I should be able to do the same, I guess.
In your line, e4 instead of Rg1 seems to be much better:

[d]1rr2bk1/3q3p/4pRp1/p2pP1P1/2pPP3/P1P3RQ/3B3P/3bN2K b - - 0 5

Maybe, black will be able to draw, I don't know.
[d]r1q2rk1/pb2bp1p/4p1p1/1p1pP1Pn/2pP4/P1P1P2Q/1PBN3P/R1B2RK1 w - - 0 17
To remind the whole move order:
17. Rf6 Qd8 18. Nf3 a5 19. Bd2 b4 20. Kh1 Bc6 21. Rg1 bxa3 22. bxa3 Rb8 23. Ne1 Qd7 24. Bd1 Ng7 25. Qh6 Ba4 26. Bg4 Rfc8 27. Rg3 Nh5 28. Bxh5 Bf8 29. Bxg6 fxg6 30. Qh3 Bd1 31. e4

Stockfish is optimistic to hold the position or even to win it after 31.e4:

[d]1rr2bk1/3q3p/4pRp1/p2pP1P1/2pPP3/P1P3RQ/3B3P/3bN2K b - - 0 31

Code: Select all

 48/85	1:03:24	139.396.477k	36.644k	+2,22	31. ... dxe4 32.Rg2 Bxa3 33.Rgf2 Bh5 34.Kg2 Rb1 35.Qe3 Bg4 36.Qxe4 Qb7 37.Qxb7 Rxb7 38.Nc2 Be7 39.R6f4 Bf5 40.Ne3 Bxg5 41.Nxf5 gxf5 42.R4f3 Bxd2 43.Rxd2 a4 44.d5 a3 45.Rff2 exd5 46.Rxd5 Ra8 47.Ra2 Rb3 48.e6 Rxc3 49.e7 Re8 50.Ra5 Rxe7 51.R5xa3 Rxa3 52.Rxa3 Kf7 53.Kf3 Kf6 54.Ra2 c3 55.Rc2 Rc7 56.Kf4 h6 57.h3 Ke6 58.Ke3 Ke5 59.Kd3 f4 60.Rxc3
Lyudmil Tsvetkov wrote:I don't know what you want to claim, that Rf6 does not win?
Certainly it does, I am just too tired now and extremely busy to do careful analysis of alternative lines.
So far I have not claimed anything, have I?
It's an very interesting and highly complex position. Rf6 is the best move and gives white advantage, but whether this advantage is enough to win is still not quite clear. Right now I would say white has good winning chances.
Lyudmil Tsvetkov wrote:Btw., after Rf6 Qd8 Nf1 Nf6 gf6 Bf6 ef6 Qf6, did you check with longer search all the lines:

[d]r4rk1/pb3p1p/4pqp1/1p1p4/2pP4/P1P1P2Q/1PB4P/R1B2NK1 w - - 0 5

I guess this is won for white, but you will need 30-40 moves until the score changes.
I myself tried this with SF and it starts winning after 30-40 moves.
From my experience drawn from different databases and top engine games, 2 minor pieces vs rook + 2 pawns, when the 2 minors also have the bishop pair is won for the minor side, in maybe at least 90% of cases, so, I suggest you

check more carefully/maybe play some games to get convinced.
I'll take again a closer look at this position.
Lyudmil Tsvetkov wrote:Alternatively, white has also Bd1 instead of Nf1:

[d]r2q1rk1/pb2bp1p/4pRp1/1p1pP1Pn/2pP4/P1P1P2Q/1P1N3P/R1BB2K1 b - - 0 2

This should also be investigated closely.

I am fully certain Rf6 wins, and maybe other moves too.
I just don't have even couple of minutes time now to give just this position.

It is a pity all the most beautiful lines are not evident from what SF chooses.
There are really some very intriguing variations here, if SF accepts the sacrifice.
Pity I don't have more time.
As I said before, it's a very complex position.
The white attack looks very dangerous. But black also has defense resources and counterplay on the queenside.

The whole thread has become quite confusing and it is becoming more and more difficult to follow the postings and analyses.
Maybe I have the time for a summary of our analyses.
Well, 32.Rg1 is certainly much better, to cover the first rank.

[d]1rr2bk1/3q3p/4pRp1/p3P1P1/2pPp3/P1P4Q/3B3P/3bN1RK b - - 0 2

I guess you and SF underestimate the power of the white central d4, e5 and g5 pawns, which, if turned passers at some point, could become very dangerous. You need long lines to see that.
I am still of the opinion white is better, but maybe the win is already not there. I should have skipped something, as said, I am just skimming through it.
I don't know if it makes sense to continue with that analysis, as this certainly could not be considered as the main line.

There is something better, I am certain about it, judging by the characteristics of the position, but I really don't have time now to investigate.

I would be curious to learn from you what your machine says about the other lines, when you have time, you might post some feedback.
Might be a bit cluttered here, but still visible.

Lyudmil Tsvetkov
Posts: 6033
Joined: Tue Jun 12, 2012 10:41 am

Re: Stockfish and Fischer

Post by Lyudmil Tsvetkov » Sat Jan 20, 2018 7:53 pm

When I am looking at it now, an even deeper and more complex alternative line is Rf6 Qd8 Nf3 a5 Qg2:

[d]r2q1rk1/1b2bp1p/4pRp1/pp1pP1Pn/2pP4/P1P1PN2/1PB3QP/R1B3K1 b - - 0 3

with the idea of h4 and Nh2-g4.

I will need to buy 128 core to tell me how this ends precisely.
Maybe this is the winning attacking line, or maybe something else I am just unable to find now quickly.

I am ready to bet 10 000 dollars the position is won for white, just don't have time to analyse now.

Maybe someone can confirm Qg2 wins?

Lyudmil Tsvetkov
Posts: 6033
Joined: Tue Jun 12, 2012 10:41 am

Re: Human versus Machine

Post by Lyudmil Tsvetkov » Fri Mar 16, 2018 11:50 am

Currently, the 1st part is just 1 dollar: https://www.amazon.com/Human-Versus-Mac ... 8&qid=&sr=
You can have the 3 for 7!
This will be just a limited offer, so don't miss it.

Lyudmil Tsvetkov
Posts: 6033
Joined: Tue Jun 12, 2012 10:41 am

Re: Human versus Machine

Post by Lyudmil Tsvetkov » Tue May 29, 2018 6:01 pm

As part of a free promotion, running for a couple of days, my latest book, 'Practical Mates', is offered for free.
Download it here: https://www.amazon.com/Practical-mates- ... B07CQVNC6K
500 mate problems, ranging from 1 to 5 moves.
You will get much better after reading it, of course.
If interested, spread the word about it.

Lyudmil Tsvetkov
Posts: 6033
Joined: Tue Jun 12, 2012 10:41 am

Re: Human versus Machine

Post by Lyudmil Tsvetkov » Thu May 31, 2018 5:40 pm

2 more days to go for the free promotion of my latest book, 'Practical Mates'.
Download it here to get stronger at checkmate pattern recognition and at chess:
https://www.amazon.com/Practical-mates- ... B07CQVNC6K

Post Reply