Total mess

[Lyudmil Tsvetkov]

[d]8/6kb/p3n1p1/Pp1nPp2/1PpPp2p/2P4P/3B2P1/2R1R1K1 b 0 1

Hi.
It is total mess on the board.

Which side would you prefer?

Best, Lyudmil

[Red Hood]

Draw. If black goes forward it help white by opening lines for white's rooks. If they penetrate, blacks pawns are gone.

[Lyudmil Tsvetkov]

Draw. If black goes forward it help white by opening lines for white's rooks. If they penetrate, blacks pawns are gone.

I do not know the outcome, but I think it is won for black, or at least 'unclear in favour of black'.

[Red Hood]

Draw. If black goes forward it help white by opening lines for white's rooks. If they penetrate, blacks pawns are gone.

I do not know the outcome, but I think it is won for black, or at least 'unclear in favour of black'.

Engines say +1.0 for black but it's only because black is already a pawn up.

I would like to see how black will this. It's a tricky position. There is no way to predict the outcome against a strong opponent like Houdini. My computer is a 2600+ sempron with 512MB of RAM so engines aren't going to give a precise evaluation on my computer unless it runs for 3 days! To beat a human with this position, sure. I mean with the clock ticking a blunder is coming. Strong pressure is there. Humans aren't good in defense.

Kind Regards, Red Hood

[Red Hood]

Two strong engines played from that position and it was a draw. After 14 moves. 3 fold repetition.

Since you wanted an engines opinion. There you go!

Kind Regards, Red Hood

[Thanar]

Here's a continuation I played out picking from among the top moves suggested by Houdini 1.5 for both sides:

1...g5 2.Kh2 f4 3.Rh1 Kf7 4.Rhd1 e3 5.Be1 Ke7 6.Ra1 Be4 7.Ra2 f3 8.gxf3 Bxf3 9.Rb1 Be4 10.Rbb2 Nef4 11.Ra1 Bf5 12.Raa2 Bxh3 13.Kg1 Ke6 14.Rh2 Bf5 15.Kf1 Bh3+ 16.Kg1 Bg4 17.Kf1 Kf5 18.Rac2 Nd3

[d]8/8/p7/Pp1nPkp1/1PpP2bp/2Pnp3/2R4R/4BK2 w - - 12 19

The game continued:

19.Ra2 Nxe1 20.Kxe1 Nxc3 21.Rag2 Nd5 22.Kf1 Nxb4 23.e6 Bh5 24.e7 c3 25.Rg1 c2 26.Kg2 Nd3 27.e8=Q Bxe8 28.Kf3 c1=Q 29.Rxc1 Nxc1 30.Kxe3 Nb3 31.Rf2+ Kg4 32.Rg2+ Kh5 33.d5 Nxa5

...with a win for black. Perhaps someone can suggest improvements for white?

[Lyudmil Tsvetkov]

Here's a continuation I played out picking from among the top moves suggested by Houdini 1.5 for both sides:

1...g5 2.Kh2 f4 3.Rh1 Kf7 4.Rhd1 e3 5.Be1 Ke7 6.Ra1 Be4 7.Ra2 f3 8.gxf3 Bxf3 9.Rb1 Be4 10.Rbb2 Nef4 11.Ra1 Bf5 12.Raa2 Bxh3 13.Kg1 Ke6 14.Rh2 Bf5 15.Kf1 Bh3+ 16.Kg1 Bg4 17.Kf1 Kf5 18.Rac2 Nd3

[d]8/8/p7/Pp1nPkp1/1PpP2bp/2Pnp3/2R4R/4BK2 w - - 12 19

The game continued:

19.Ra2 Nxe1 20.Kxe1 Nxc3 21.Rag2 Nd5 22.Kf1 Nxb4 23.e6 Bh5 24.e7 c3 25.Rg1 c2 26.Kg2 Nd3 27.e8=Q Bxe8 28.Kf3 c1=Q 29.Rxc1 Nxc1 30.Kxe3 Nb3 31.Rf2+ Kg4 32.Rg2+ Kh5 33.d5 Nxa5

...with a win for black. Perhaps someone can suggest improvements for white?

Hi Terry.
Many thanks for your effort and the interesting analysis. Houdini 1.5 plays it fantastically well, I did not think of many moves suggested by it. But engines are definitely better in such positions with many fronts and many attacked objects. The white pawns are falling one after the other, and maybe this is the effect of the enormous black space advantage and the power of the black chain.

Could you please shed some light on what are the evals of Houdini? At what point does the engine start seeing black in advantage and what about the initial position? Seemingly white has big material advantage, does Houdini see it all due to the search?

Many thanks again for the effort.

Best, Lyudmil

[Lyudmil Tsvetkov]

Two strong engines played from that position and it was a draw. After 14 moves. 3 fold repetition.

Since you wanted an engines opinion. There you go!

Kind Regards, Red Hood

Hi Milan.
Many thanks for investing time in analysis!
But you see that Terry has found a winning continuation with the aid of Houdini 1.5.
On my part, as I am much more naive than Houdini, and do not notice many strong continuations based on tactics, at least with limited thinking time, I would go for a forced line leading forcefully to this position:

[d]8/8/p3n3/Pp1nP1p1/1PpPkp1p/2Pbp2P/R5P1/2R1B1K1 b 0 1

and then play 1...g4 2. hg4 h3 3. gh3 f3

[d]8/8/p3n3/Pp1nP3/1PpPk1P1/2Pbpp1P/R7/2R1B1K1 w 0 1

probably losing the game in the end, as it is possible that some of the black advantage has already slipped away, and I have headache just looking at the last position: for a final verdict I would need at least half a week... But then, engines are much better here, maybe they would know better.

The simple matter of fact my humble and limited human knowledge presents me with in the initial position is that white does not have counterplay, so it is a one-sided affair, assuring black of at least a draw. But then, playing it is difficult.

Many thanks for contributing.

Best, Lyudmil

[Thanar]

Could you please shed some light on what are the evals of Houdini? At what point does the engine start seeing black in advantage and what about the initial position?

Houdini 1.5’s evaluation of the initial position is in White’s favour, with the evaluation generally sinking as it has more time to evaluate. For example at depth 20, its evaluation is +1.00, but it drops to +.73 at depth 21.

Here are some Houdini 1.5 evaluations for my continuation (1...g5 2.Kh2 f4 3.Rh1 Kf7 4.Rhd1 e3 5.Be1 Ke7 6.Ra1 Be4 7.Ra2 f3 8.gxf3 Bxf3 9.Rb1 Be4 10.Rbb2 Nef4 11.Ra1 Bf5 12.Raa2 Bxh3 13.Kg1 Ke6 14.Rh2 Bf5 15.Kf1 Bh3+ 16.Kg1 Bg4 17.Kf1 Kf5 18.Rac2 Nd3 19.Ra2 Nxe1 20.Kxe1 Nxc3 21.Rag2 Nd5 22.Kf1 Nxb4 23.e6 Bh5 24.e7 c3 25.Rg1 c2 26.Kg2 Nd3 27.e8=Q Bxe8 28.Kf3 c1=Q 29.Rxc1 Nxc1 30.Kxe3 Nb3 31.Rf2+ Kg4 32.Rg2+ Kh5 33.d5 Nxa5):

From 4…e3 to 6.Ra1, Houdini 1.5 evaluates the position as equal (0.0 at depth 21). But after Be4, the evaluation tips slightly in Black’s favour (-0.22 at depth 25). The evaluation remains at -0.22 until 11…Bf5 where it moves to -.80 at depth 20 (and plummets further as the search depth increases). After 18...Nd3, the evaluation is -1.26 at depth 19, and after 20...Nxc3 it is -3.12 at depth 17.

[Thanar]

Here is another very interesting continuation based on Houdini 1.5 analysis:
Complete pgn for cutting & pasting:

1...g5 2.Rcd1 f4 3.Re2 3...Kf7 4.Rc1 e3 5.Be1 f3 6.gxf3 Nef4 7.Rh2 Bf5 8.Kh1 Bxh3 {eval 0.00 at depth 23} 9.Rcc2 Bf5 10.Ra2 Ke6 11.Ra1 Bd3 {eval -.47 at depth 20} 12.Bxh4 gxh4 13.Rxh4 Ng6 14.Rh7 Nxe5 15.Rh6+ Ng6 16.Re1 Nxc3 17.Rxe3+ Kd5 18.Rh2 Nf4 19.Re1 Be2 20.Kg1 Kxd4 21.Rh6 Nd3 22.Ra1 Nxb4 23.Kf2 Nd3+ 24.Kg3 b4 {eval -2.09 at depth 16} 25.Rah1 b3 26.Rxa6 b2 27.Rd6+ Ke3 28.Re6+ Kd2 29.a6 Nb5 30.Rb6 {eval -5.57 at depth 18} Na7 31.f4 c3 32.Rb1 Bd1 33.f5 Kc2 34.R1xb2+ cxb2 35.Kh4 Kc3 36.Rxb2 Nxb2 37.f6 Bb3 38.Kg5 Nd3 39.Kg6 Ne5+ {eval -9.96 at depth 18, mate in 20 at depth 24}

1...g5 2.Rcd1 f4 3.Re2 3...Kf7 4.Rc1 e3 5.Be1 f3 6.gxf3 Nef4 7.Rh2 Bf5 8.Kh1 Bxh3 {eval 0.00 at depth 23} 9.Rcc2 Bf5 10.Ra2 Ke6 11.Ra1 Bd3 {eval -.47 at depth 20}

[d]8/8/p3k3/Pp1nP1p1/1PpP1n1p/2PbpP2/7R/R3B2K w - - 6 12
White now sacrifices a bishop to eliminate Black’s kingside pawns:

12.Bxh4 gxh4 13.Rxh4 Ng6 14.Rh7 Nxe5 15.Rh6+ Ng6 16.Re1 Nxc3 17.Rxe3+ Kd5 18.Rh2 Nf4 19.Re1 Be2 20.Kg1 Kxd4 {eval -2.25 at depth 17}

[d]8/8/p7/Pp6/1Ppk1n2/2n2P2/4b2R/4R1K1 w - - 0 21
Can white's two rooks achieve enough counterplay and perhaps get some white pawns moving?

21.Rh6 Nd3 22.Ra1 Nxb4 23.Kf2 Nd3+ 24.Kg3 b4 {eval -2.09 at depth 16}

[d]8/8/p6R/P7/1ppk4/2nn1PK1/4b3/R7 w - - 0 25
The answer is no! Black blocked the rooks from the queenside and has formed a moving cluster of his pieces and King to escort the queenside(!) pawns to victory:

25.Rah1 b3 26.Rxa6 b2 27.Rd6+ Ke3 28.Re6+ Kd2 29.a6 Nb5 30.Rb6 {eval -5.57 at depth 18} Na7 31.f4 c3 32.Rb1 Bd1 33.f5 Kc2 34.R1xb2+ cxb2 35.Kh4 Kc3 36.Rxb2 Nxb2 37.f6 Bb3 38.Kg5 Nd3 39.Kg6 Ne5+ {eval -9.96 at depth 18, claim of mate in 20 at depth 24}

[d]8/n7/P4PK1/4n3/8/1bk5/8/8 w - - 5 40