Is this position a win for white?

[MikeGL]

I also tried, and it seems h4 instead of Nxb6 could win for white.

Sorry, I tried again. Defence of black remains the same and I can not find the win.

Ok, my third attempt thanks to good ideas presented by other posters above.
Looks like a 1-0 for white with Nc6 invasion for the win.

Analysis was done with engine assistance and ideas of above posters.
Maybe a forced win, or the defence I played is not accurate.

[pgn]

[Event "?"]
[Site "?"]
[Date "2018.01.29"]
[Round "?"]
[White "?"]
[Black "Stockfish8"]
[Result "1-0"]
[BlackElo "2400"]
[WhiteElo "2400"]
[TimeControl "1800+10"]
[SetUp "1"]
[FEN "2k5/2p5/1q1p4/pPpPp1pp/N1P1Pp2/P4PbP/KQ4P1/8 w - - 0 1"]
[Termination "unterminated"]
[PlyCount "28"]
[WhiteType "human"]
[BlackType "human"]

1. Qd2 Qa7 2. Kb3 Bh2 3. Nb2 Bg3 4. Ka4 Kb7 5. Nd3 Bh4 6. Nc1 Qa8 7. Nb3
Qa7 8. Qxa5 Qb6 (8. .. Qxa5+ 9. Nxa5+ Kc8 10. Nc6 Kd7 11. Ka5+-) 9. Qd2
Qa7+ 10. Qa5 Qb6 11. Qa6+ Qxa6+ 12. bxa6+ Kxa6 13. Na5 Kb6 14. Nc6 1-0
[/pgn]

What if 7...Bg3 and if 8. Qxa5 then Be1? Big difference between having the Q on a7 and having it on a8.

The line you provided and the one by Uri seems to hold.
Quite embarassing, i thought Nc6 position can be reached.
Both your lines above stops the idea.

[BeyondCritics]

I think i have figured out some serious winning attempts.
My first experiments were unfortunately hampered by bad commentary from the pre-stockfish era.
First: 0...Qb6 was not a joke. It prevents the breakthrough b6. And second, the idea to gang on the a5 pawn is much to simple minded, since black has some serious counterplay on the dark squares (a4, Bd4) and with g4 in some lines.
The proposed winning plan runs as follows:

• -Leave Queen and Night at their active places and activate your king. Prevent the activation of the black bishop at all costs. If Black misbehaves, punish him with b6 or Nxc5! or Nxb6.
  -When the white king arrives on e2, restricting the black bishop, put the queen on b1, threatening Nxb6. This forces black to move his queen away from b6.
  -If needed, disorganize his forces somehow, using the principle of the two weaknesses.
  -Break through with b6 and put your queen on b5, forcing him to trade on b5. Bring your knight to c4. This hopefully leads to a winning endgame.

Some example lines:
[pgn]
[Event "Nxb6 Trap"]
[Site ""]
[Date "2018.01.29"]
[Round ""]
[White ""]
[Black ""]
[Result "*"]
[Annotator "CFish 290118, O.Roese"]
[FEN "2k5/2p5/1q1p4/pPpPp1pp/N1P1Pp2/P4PbP/KQ4P1/8 w - - 0 1"]

1.Kb1 $1 Kb8
( 1...Qb7 2.Kc1
( 2.Nxc5 $2 Qb6 $1 $10 )
( 2.Kc2 $2 Bf2 )
( 2.b6 $2 cxb6 3.Qb5 Kd8 4.Nxb6 Kc7 5.Na4 Bf2 6.Kc2 Qxb5 7.cxb5
Bd4 $10 )
2...Qb6
( 2...Kd8 $4 3.Nxc5 dxc5 4.Qxe5 Qc8 5.Qxg5+ $20 )
( 2...Kb8 $2 3.b6 $18 )
( 2...Kd7 $4 3.Nxc5+ $20 )
( 2...Bh2 3.b6 $18 )
( 2...Bh4 3.b6 cxb6 4.Qb5 Kd8 5.Nxb6 Be1 6.Na4 Qxb5 7.cxb5 $18 )
( 2...Be1 3.b6 cxb6 4.Qb5 Kd8 5.Nxb6 Ke7 6.Kc2 $18 )
3.Kd1 )
( 1...Qb8 2.b6 Qb7 3.Qb5 Kd8 4.bxc7+ $1
( 4.Qxa5 $2 cxb6 5.Qb5 Bf2 6.Nb2 Be1 7.Nd3 Ba5 8.Nc1 Qd7 9.Nb3 Kc8
$10 )
4...Kxc7 5.Kc2 Qxb5 6.cxb5 Bf2 7.Nb2 $16 c4 8.Nxc4 a4 9.Kc3 Be1+ 10.
Kb2 Bf2 11.Ka2 Bd4 12.b6+ Kb8 13.Nxd6 Bxb6 14.Nf7 Bd4 )
( 1...Be1 $4 2.Nxb6+ cxb6 3.h4 $20 )
( 1...Bh2 $4 2.Nxb6+ cxb6 3.h4 gxh4 4.g3 $20 )
2.Kc1 Kc8 3.Kd1 Qb7
( 3...Kb8 4.Ke2 Kc8 5.Qb1 Qb7
( 5...Kb8 6.Nxb6 cxb6 7.h4 $20 )
6.b6 cxb6
( 6...c6 7.Qd1 $18 )
7.Qb5 Kd8 8.Nxb6 Kc7 9.Na4 Qxb5 10.cxb5 $18 )
4.Ke2 Bh2 5.Qb1 Kd8 6.Qe1 $1
( 6.b6 $6 cxb6 7.Qb5 Bg1 8.Nxb6 Kc7 9.Na4 Qxb5 10.cxb5 Bd4 $10 )
6...Qa8 7.Qh1 Bg3 8.Qb1 Bh2 9.b6 Qb7 10.bxc7+ Kxc7 11.Qb5 Qxb5 12.cxb5 c4
13.Nb2 $16 *
[/pgn]
There are several fortress like endgames here, which stockfish/cfish does not like, so always check the pv for plausibility.

[Eelco de Groot]

Maybe we can look at the position again, with some improved programming in more than a year and a half and some people with very big machines ? Maybe see new things as requested also here http://www.talkchess.com/forum3/viewtop ... art=70#top

The questions still pending for these positions are :

How to win this position with white ?
2k5/2p5/1q1p4/pPpPp1pp/N1P1Pp2/P4PbP/KQ4P1/8 w - - am Nxb6+?; c1 "diff=10";

I took Kaissa and Kaissa or Stockfish does not agree with all of the pgn in above post, but I picked the first move anyway and with Kaissa I get after 1. Kb1


[d]2k5/2p5/1q1p4/pPpPp1pp/N1P1Pp2/P4PbP/1Q4P1/1K6 b - -

Engine: Kaissa IV NoContempt (1 thread, no tablebases, 512 MB)
by T. Romstad, M. Costalba, J. Kiiski, G. Linscott

49 773:34 +2.76 1...Qb7 2.Ka2 Be1 3.b6 cxb6 4.Qb5 Kd8
5.Nxb6 Bd2 6.Na4 Qc7 7.Qa6 Qb8 8.Qc6 Qc7
9.Qb5 Be1 10.Nb6 Qb7 11.Kb3 Bf2
12.Na4 Qc7 13.Qa6 Qb8+ 14.Kc2 Be1 (93.784.976.048) 2020

49 773:34 +3.40 1...Qb8 2.b6 cxb6 3.Qb5 Kd8 4.Nxb6 Qb7
5.Na4 Qc7 6.Qa6 Be1 7.Nb6 Qb8 8.Kc1 Bf2
9.Qxa5 Qb7 10.Na4+ Ke7 11.Qb5 Qa7
12.Kc2 g4 13.Kb3 gxf3 14.gxf3 Be1 (93.784.976.048) 2020

49 773:34 +3.81 1...Qa7 2.b6 cxb6 3.Qb5 Kd8 4.Nxb6 a4
5.Qc6 Qc7 6.Qa8+ Ke7 7.Nxa4 Be1
8.Qg8 Qb7+ 9.Nb2 Qa6 10.Qxg5+ Kd7
11.Qg7+ Kc8 12.Qh8+ Kd7 13.Qh7+ Kd8
14.Qxh5 Bc3 (93.784.976.048) 2020

49 773:34 +8.05 1...h4 2.Nxb6+ cxb6 3.Qe2 Kb7 4.Kc2 Kb8
5.Qf1 Kc8 6.Kd3 Kb7 7.Kc3 Kc8 8.Qb1 Kb7
9.Kc2 Kb8 10.Kc1 Kc7 11.Qa1 Kc8
12.a4 Kb8 13.Kd2 Kb7 14.Ke2 Kc7 (93.784.976.048) 2020

Here Black's second reply in top line is 2...Be1, before that it was 2...Bh4 which maybe a little less in value but that was the move I looked at deeper first. I assume 1...Qb7 is best, at least for now and see if the score for White can still be improved a bit.

This gives at the moment, in a new instance of Kaissa, after 1. Kb1 Qb7 2. Ka2


[d]2k5/1qp5/3p4/pPpPp1pp/N1P1Pp2/P4PbP/KQ4P1/8 b - -

Engine: Kaissa IV NoContempt (512 MB)
by T. Romstad, M. Costalba, J. Kiiski, G. Linscott

50 638:44 +3.30 2...Be1 3.b6 cxb6 4.Qb5 Kd8 5.Nxb6 Bd2
6.Na4 Qa7 7.Kb3 Qc7 8.Kb2 Qa7 9.Kc2 Be1
10.Kc1 Ke7 11.Kb2 Kd8 12.Nb6 a4
13.Nxa4 Qa5 14.Qb8+ Ke7 15.Kb3 Qd2 (78.964.628.214) 2060

50 638:44 +3.37 2...Bh4 3.b6 Kd8 4.Qb5 cxb6 5.Nxb6 Be1
6.Na4 Qc7 7.Kb3 Bd2 8.Kb2 Qa7 9.Kc2 Be1
10.Kc1 Ke7 11.Kb2 Kd8 12.Nb6 a4
13.Nxa4 Qa5 14.Kb3 Qd2 15.Qb8+ Ke7 (78.964.628.214) 2060

50 638:44 +3.69 2...Bh2 3.b6 cxb6 4.Qb5 Kd8 5.Nxb6 Kc7
6.Na4 Qxb5 7.cxb5 c4 8.Nb2 c3 9.Nc4 a4
10.b6+ Kb7 11.Nxd6+ Kxb6 12.Nf7 Bg1
13.Kb1 Bd4 14.Kc2 Kc7 15.Nxg5 Kc8 (78.964.628.214) 2060

49 638:44 +3.51 2...Qa7 3.b6 cxb6 4.Qb5 Kd8 5.Nxb6 Qb7
6.Na4 Qc7 7.Qa6 Be1 8.Kb3 Qb8+ 9.Kc2 Ke7
10.Nb6 Kd8 11.Na8 a4 12.Nb6 Kc7
13.Nxa4 Qb7 14.Qb5 Qa8 15.Qb6+ Kc8 (78.964.628.214) 2060

which is not a bad score I think

(I do not really know how to interpret the high score for White after

49 773:34 +8.05 1...h4 2.Nxb6+ cxb6 3.Qe2 Kb7 4.Kc2 Kb8
5.Qf1 Kc8 6.Kd3 Kb7 7.Kc3 Kc8 8.Qb1 Kb7
9.Kc2 Kb8 10.Kc1 Kc7 11.Qa1 Kc8
12.a4 Kb8 13.Kd2 Kb7 14.Ke2 Kc7 (93.784.976.048) 2020

Hard to believe that + 8.05 would still be a draw and Queen then can't be taken here either but not 100% sure it wins.
Score has not gone up in a long while.)