Is this a fortress?

[glav]

I have been analyzing the following position using stockfish. Despite the fact that a considerable plus score is given to white, the program seems to have no idea of what is going on. Indeed, if black refrains from pushing in b6 or in b5 or in h5 and guards agains the push to b5, white keeps shuffling the pieces back and forth, without making any real progress. I was wondering if things could improve using diferent search parameters or different engines.
Thanks in advance.
[pgn]
[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[FEN "r3b1k1/1p2q1pp/p1p1pr2/2PpN3/PP1PpPP1/4P3/3Q3P/R4RK1 w - - 0 25"]

25. Qe1 Bd7 26. Qh4 Be8 27. Kg2 Bd7 28. Rf2 Rb8 29. Rc2 Be8 30. Rb1 Bd7
31. Kg1 Be8 32. Rcb2 Qc7 33. Qg3 Qe7 34. h3 Rc8 35. Kh1 Rb8 36. Kh2 Kh8
37. Kg1 Rc8 38. h4 Kg8 39. h5 Rb8 40. Kg2 Rf8 41. Kh3 Qc7 42. Rc1 Qd8 43.
Rh1 Qe7 44. Kg2 Rf6 45. Qh4 Qd8 46. Rbb1 Ra8 47. Rbf1 Rc8 48. Qg3 Rc7 49.
Rh3 Qe7 50. Rh2 Qf8 51. Rfh1 Rc8 52. Qh4 Ra8 53. Ra1 *
[/pgn]

[Graham Banks]

I don't think it's a fortress.

[glav]

I don't think it's a fortress.

Then, how could White make progress?

[peter]

I have been analyzing the following position using stockfish. Despite the fact that a considerable plus score is given to white, the program seems to have no idea of what is going on. Indeed, if black refrains from pushing in b6 or in b5 or in h5 and guards agains the push to b5, white keeps shuffling the pieces back and forth, without making any real progress. I was wondering if things could improve using diferent search parameters or different engines.
Thanks in advance.
[pgn]
[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "*"]
[FEN "r3b1k1/1p2q1pp/p1p1pr2/2PpN3/PP1PpPP1/4P3/3Q3P/R4RK1 w - - 0 25"]

25. Qe1 Bd7 26. Qh4 Be8 27. Kg2 Bd7 28. Rf2 Rb8 29. Rc2 Be8 30. Rb1 Bd7
31. Kg1 Be8 32. Rcb2 Qc7 33. Qg3 Qe7 34. h3 Rc8 35. Kh1 Rb8 36. Kh2 Kh8
37. Kg1 Rc8 38. h4 Kg8 39. h5 Rb8 40. Kg2 Rf8 41. Kh3 Qc7 42. Rc1 Qd8 43.
Rh1 Qe7 44. Kg2 Rf6 45. Qh4 Qd8 46. Rbb1 Ra8 47. Rbf1 Rc8 48. Qg3 Rc7 49.
Rh3 Qe7 50. Rh2 Qf8 51. Rfh1 Rc8 52. Qh4 Ra8 53. Ra1 *
[/pgn]

LC0 with net 3905000 at 25'+5" played out on 3070ti like this:

[pgn][Event "Stellung ausspielen (Lcv3-dag+git1842e1"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "?"]
[Black "?"]
[Result "1-0"]
[SetUp "1"]
[FEN "r3b1k1/1p2q1pp/p1p1pr2/2PpN3/PP1PpPP1/4P3/3Q3P/R4RK1 w - - 0 25"]
[PlyCount "25"]

25. Qe1 {4.76/6} Rf8 {4.44/8} 26. Qg3 {4.43/8} Rb8 {4.03/8} 27. Rab1 {4.40/9} Qc7 {4.05/9} 28. Kg2 {4.58/9} b6 {4.26/9} 29. Qe1 {4.92/10} Rb7 {4.50/9} 30. Qc3 {5.17/10} Qd8 {4.81/10} 31. b5 {5.11/14 (Kg3)} cxb5 {4.89/16} 32. c6 {5.27/16} Rc7 {5.17/14} 33. axb5 {5.64/15} axb5 {5.46/15} 34. Rxb5 {6.27/15} Rc8 {6.21/14} 35. Rb2 {7.12/13} h5 {6.36/14 (g5)} 36. g5 {7.00/18} Rf5 {6.52/13} 37. Qb4 {7.85/13} 1-0[/pgn]

Here GUI adjudicated as 1-0, anyhow progess is clearly to be seen after White's b5, isn't it? Of course Black's ...b6 wasn't necessary, but not doing that wouldn't stop White's advance of pawns neither, regards

[glav]

Thanks for your reply. Your variation shows a clear progress, but, as I said in my post, B should avoid playing pushes in b6 (like your 28 move) or in b5 or in h5. In these cases, just play the second best. By doing this and by taking profylaxis against the W push to b5, I could not see any progress,

[peter]

Thanks for your reply. Your variation shows a clear progress, but, as I said in my post, B should avoid playing pushes in b6 (like your 28 move) or in b5 or in h5. In these cases, just play the second best. By doing this and by taking profylaxis against the W push to b5, I could not see any progress,

Just hop in at LC0's game with any alternative to 28...b6, if you like to, regards

[glav]

Thanks for your reply. Your variation shows a clear progress, but, as I said in my post, B should avoid playing pushes in b6 (like your 28 move) or in b5 or in h5. In these cases, just play the second best. By doing this and by taking profylaxis against the W push to b5, I could not see any progress,

Just hop in at LC0's game with any alternative to 28...b6, if you like to, regards

Unfortunately, I don't have access to any decent machine with LC0. However, instead of 28)...b6, it seems to me that other moves like 28)..Ra8 could be safely played without harming the fortress.

[peter]

Thanks for your reply. Your variation shows a clear progress, but, as I said in my post, B should avoid playing pushes in b6 (like your 28 move) or in b5 or in h5. In these cases, just play the second best. By doing this and by taking profylaxis against the W push to b5, I could not see any progress,

Just hop in at LC0's game with any alternative to 28...b6, if you like to, regards

Unfortunately, I don't have access to any decent machine with LC0. However, instead of 28)...b6, it seems to me that other moves like 28)..Ra8 could be safely played without harming the fortress.

LC0 after 28...Ra8 played out against SF dev. with 30 threads of 16x3.5GHz, 8G hash and same TC of 25'+5" like this:

[pgn][Event "AMD, Schnellschach 25.0min+5.0sek"]
[Site "Dort"]
[Date "2023.05.10"]
[Round "1"]
[White "Lc0 v0.30.0-dag+git.1842e13d"]
[Black "Stockfish"]
[Result "1-0"]
[Annotator "5.29;2.14"]
[SetUp "1"]
[FEN "r3b1k1/1p2q1pp/p1p1pr2/2PpN3/PP1PpPP1/4P3/3Q3P/R4RK1 w - - 0 25"]
[PlyCount "33"]
[TimeControl "1500+5"]

25. Qe1 Rf8 26. Qg3 Rb8 27. Rab1 Qc7 28. Kg2 Ra8 29. Rg1 {5.29/8 74} Rc8 {2.14/59 209 (De7)} 30. h4 {6.20/10 20} Rf6 {2.68/48 179 (h5)} 31. h5 {6.21/9 57} Qd8 {3.10/45 133 (De7)} 32. Rb2 {5.97/8 44 (Th1)} Qe7 {3.00/40 64} 33. Qh4 {5.74/8 34 (Th1)} Rb8 {2.95/40 80 (Tc7)} 34. h6 {8.40/9 18 (Th1)} b6 {3.54/42 186 (gxh6)} 35. cxb6 {8.10/12 37 (Tbb1)} Rxb6 {2.97/32 15} 36. hxg7 {8.71/11 1} Qxg7 {3.67/37 173 (Tf7)} 37. g5 {9.56/10 48 (Kg3)} Rf5 {3.05/28 12} 38. Kh1 {9.40/9 47} Rb7 {3.63/37 108} 39. Rc1 {8.18/10 42 (Dh3)} Rc7 {3.80/33 42 (h6)} 40. Qh3 {10.09/9 53} Qe7 {4.04/37 38 (Tb7)} 41. Qh6 {12.60/8 21 adjud.} 1-0
[/pgn]

Guess your next deviation would be from 30...Rf6, because after 34.h6 "fortress" is broken already again, isn't it?
Anyhow I just don't think, Black can stop all of White's pawns together at all, b, h, g and f, just by shuffling only without pawn- moves of its own, regards

[glav]

Thanks for your careful analysis, Peter. I learned from it another element of this position. The "criptic" move Qh4 actually threatens the push 34)h6, takind advantage that the Black rook is pinned by the Queen and, therefore, cannot take on h6. Therefore the move 33)...Kf8 or 33)... Qc7 (instead 33)... Rb8) should be played, both moves unpinning the rook. Does it still hold?

[peter]

Thanks for your careful analysis, Peter. I learned from it another element of this position. The "criptic" move Qh4 actually threatens the push 34)h6, takind advantage that the Black rook is pinned by the Queen and, therefore, cannot take on h6. Therefore the move 33)...Kf8 or 33)... Qc7 (instead 33)... Rb8) should be played, both moves unpinning the rook. Does it still hold?

In meantime I had another one game, this time I did let LC0 play against Crystal with same hardware- TC as before (against SF dev) from very starting postion again:
[pgn][Event "AMD, Schnellschach 25.0min+5.0sek"]
[Site "Dort"]
[Date "2023.05.10"]
[Round "1"]
[White "Lc0 v0.30.0-dag+git.1842e13d"]
[Black "Crystal 5 KWK"]
[Result "1-0"]
[Annotator "4.71;3.10"]
[SetUp "1"]
[FEN "r3b1k1/1p2q1pp/p1p1pr2/2PpN3/PP1PpPP1/4P3/3Q3P/R4RK1 w - - 0 25"]
[PlyCount "93"]
[TimeControl "1500+5"]

25. Qg2 {4.71/6 17} Kh8 {3.10/37 143} 26. h4 {4.78/10 23 (Dg3)} Kg8 {3.29/34 148 (h5)} 27. h5 {7.89/8 18 (Dg3)} Bd7 {3.97/35 149} 28. Kh2 {7.00/8 60 (Dg3)} Be8 {4.22/32 118} 29. Qh3 {6.43/8 31} Qc7 {4.14/33 75 (Ld7)} 30. Ra2 {5.89/8 36 (Tf2)} Bd7 {4.18/33 148 (De7)} 31. Qh4 {5.35/8 21} Raf8 {4.72/35 119 (Le8)} 32. Kg3 {5.79/9 39 (Kh1)} Be8 {5.25/33 133} 33. Rh1 {5.41/11 35 (Th2)} Qb8 {5.01/29 34 (Kh8)} 34. Qg5 {5.49/10 24 (Tb1)} h6 {5.27/30 44 (Dc7)} 35. Qh4 {7.68/9 19} Kh7 {5.41/30 36 (Dc7)} 36. Rg2 {7.34/9 40 (Kh2)} b6 {5.83/30 126 (Dc7)} 37. Kh2 {9.05/9 28} Rg8 {6.20/31 58 (Db7)} 38. Rhg1 {7.90/11 63 (Tc1)} Qb7 {6.20/31 90 (bxc5)} 39. Kh1 {8.75/12 30 (Tc1)} a5 {6.45/27 33 (bxc5)} 40. Qe1 {13.06/10 19} bxc5 {6.82/30 66} 41. bxc5 {14.52/10 10} Qb3 {7.18/30 40} 42. Re2 {15.74/10 75 (Th2)} Qb7 {7.97/26 22 (Dxa4)} 43. Qxa5 {22.74/8 23 (Db1)} Qb3 {7.72/23 8} 44. Qa7 {15.00/9 48 (Dd2)} Kh8 {7.40/19 2 (Tgf8)} 45. a5 {16.18/8 24} Qb5 {8.92/25 0} 46. Rge1 {22.38/9 26} Bxh5 {9.54/26 0 (Tff8)} 47. gxh5 {25.70/8 33 (Db6)} Kh7 {11.19/26 17 (Tf5)} 48. Qb6 {30.73/8 53} Qa4 {14.58/27 0 (Tf5)} 49. a6 {27.92/8 33 (Tb1)} Rf5 {9.68/23 4} 50. Qxc6 {34.79/8 28 (Th2)} Rxh5+ {10.44/18 3} 51. Kg1 {35.03/9 33} Qxc6 {14.38/24 0} 52. Nxc6 {60.30/8 11} Ra8 {16.13/22 7} 53. a7 {61.33/7 55 (Ta1)} Rf5 {18.00/21 16} 54. Rb1 {62.46/7 42} e5 {30.28/21 0 (Tf7)} 55. dxe5 {80.95/6 33 (fxe5)} d4 {16.62/21 10 (Tf7)} 56. Rb7 {59.47/6 36 (Sxd4)} d3 {33.92/21 7 (dxe3)} 57. Rg2 {33.74/7 24} d2 {#19/42 0} 58. Rxd2 {61.51/7 23} Rff8 {#11/63 0} 59. Rg2 {38.19/6 53 (Tdd7)} Rg8 {#11/58 3} 60. Ne7 {43.35/7 30} Rgd8 {#10/81 0 (Txa7)} 61. f5 {44.87/7 35 (Sf5)} Rd1+ {#10/63 5 (Txa7)} 62. Kh2 {42.02/7 19} Rb1 {#8/89 0} 63. Rxb1 {78.75/7 31 (c6)} Rxa7 {#7/97 6} 64. f6 {71.46/6 40 (Sg6)} g5 {#7/76 5} 65. Rb8 {56.47/6 36 (Txg5)} Rd7 {#6/86 6 (Tc7)} 66. Rg8 {#6/1 0 (f7)} Rxe7 {#5/129 5} 67. fxe7 {#5/1 0} Kxg8 {#4/182 5} 68. e8=Q+ {#4/1 0} Kg7 {#3/244 0} 69. Rd2 {#3/1 0 (Tb2)} h5 {#2/242 26} 70. Rd7+ {#2/1 0} Kh6 {#1/244 0} 71. Qe6# {#1/1 0} 1-0
[/pgn]
Will give your 33...Kf8 a last one try LC0- Crystal and return when it's finished again.

Already seen enough in edit- time, because Crystal again has to have pawn- moves again very soon:

[pgn][Event "AMD, Schnellschach 25.0min+5.0sek"]
[Site "Dort"]
[Date "2023.05.10"]
[Round "1"]
[White "Lc0 v0.30.0-dag+git.1842e13d"]
[Black "Crystal 5 KWK"]
[Result "1-0"]
[Annotator "6.87;4.50"]
[SetUp "1"]
[FEN "r3b1k1/1p2q1pp/p1p1pr2/2PpN3/PP1PpPP1/4P3/3Q3P/R4RK1 w - - 0 25"]
[PlyCount "25"]
[TimeControl "1500+5"]

25. Qe1 Rf8 26. Qg3 Rb8 27. Rab1 Qc7 28. Kg2 Ra8 29. Rg1 Rc8 30. h4 Rf6 31. h5 Qd8 32. Rb2 Qe7 33. Qh4 Kf8 34. Qg5 {6.87/9 113} Rb8 {4.50/39 288 (Th6)} 35. Rh1 {6.29/12 42 (Dh4)} h6 {6.02/35 128 (Th6)} 36. Qh4 {8.29/10 28} b6 {6.87/35 185} 37. Qf2 {9.68/11 67 (Tbb1) adjud.} 1-0
[/pgn]
So I stopped that again and did Crystal have some output at the position before 35...h6 in MultiPV=2- mode:

[fen]1r2bk2/1p2q1pp/p1p1pr2/2PpN1QP/PP1PpPP1/4P3/1R4K1/7R b - - 0 1[/fen]

Analysis by Crystal 5 KWK:

1. +- (6.03): 35...h6 36.Dh4 b6 37.Tbb1 Kg8 38.a5 bxc5 39.bxc5 Tb5 40.Txb5 cxb5 41.Tc1 Tf7 42.Dg3 Tf8 43.Kh3 Ld7 44.Sg6 De8 45.c6 Lc8 46.Sxf8 Kxf8 47.c7 Kf7 48.f5 exf5 49.Dd6 De7 50.Dxe7+ Kxe7 51.gxf5 Lxf5+ 52.Kg3 Lc8 53.Kf4 Kd6 54.Tg1 b4
2. +- (6.95): 35...Ta8 36.h6 g6 37.Tf1 Tc8 38.Ta2 Td8 39.f5 Txf5 40.Dxe7+ Kxe7 41.gxf5 exf5 42.Kh2 Ta8 43.Tg2 Kf6 44.Tfg1 Ke6 45.Ta1 Kf6 46.b5 cxb5 47.axb5 Lxb5 48.Tag1 Le8 49.c6 bxc6 50.Tb1 Ta7 51.Tgb2 c5 52.Tb6+ Kg5 53.Tg1+ Kh4 54.dxc5
Weiß steht klar auf Gewinn

As you see, second best instead of ...h6 has 36.h6 for White and eval near to +7, regards