Fortress Positions

[chrisw]

Does anyone have, or can point to any voluminous EPD suites of fortress positions? I have about one hundred and fifty on my github page, but need more, as ever. Thanks in advance.

[Jouni]

I looked at 169 position EPD. There were only 20-30 fortress positions. Most positions very familiar tactical shots.

[peter]

Indeed the fortresses.epd there seems to be a rather inhomogeneous one, best moves missing to a big part too.

Problem with fortresses even more than with other tactical hard positions is, the margin between too easy and too hard for engines is even smaller. In big databases like HHdb eg., because these being composed for human players, these are to >99% too easy for modern engines, otherwise you could just take almost all the studies with claim "...to move and draw", getting fortresses that way plenty of at once. On the other hand, really (for engines) difficult ones to a big part are still too hard for STC engine- testing without Forward- Backward at all.
Here I have 22 examples of not much too easy and not much too hard ones filtered out of one mixed database of mine with first move only and some mimimum of sources given, checked for not having duplicates to fortresses.epd with epdutil:

Code: Select all

1N4K1/1pp1p3/7k/2r3pP/p2pp1P1/P6N/8/2B5 w - - bm Nc6; hmvc 0; id "Mihai Neghina_Standstill_Dort, 1";
8/8/3kp1pp/p2p1p2/Pp1P1P1P/1K4P1/8/8 w - - bm Kc2; hmvc 0; id "fort, ress_?_CCC, 2";
1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - bm Rxh1; hmvc 0; id "eighth, step_to, heaven_?, 3";
K2b2r1/8/3p2p1/1Nk5/4Q3/p2P4/Pp2R1p1/2r2b2 w - - bm Kb7; hmvc 0; id "Krug, Peter_Stadlhofer, Anton_CSS, 4";
8/1p6/1Pp2N1q/p1Ppk2p/P3p3/3PPpPp/3K1P1P/1R6 w - - bm Ng4+; hmvc 0; id "Shredder, Chiron_Critter, nicht_Hasek 1951, 5";
4B3/k1K1pp2/p3bq2/2p1p3/1p2P1P1/4P2Q/nR6/8 w - - bm g5; hmvc 0; id "Neghina, Mihai_CSS, Forum_Selbstpatt, 6";
7B/7P/8/2p5/p1P2N2/P2p2pp/p1PPpnp1/2K1kbrq w - - bm Ba1; hmvc 0; id "Ziak=F_(=3344.58c1e1)_Schach Aktiv, Gull, 7";
7Q/4p3/4p3/p1p1P3/Pp2P3/3Kp3/p1PbP3/kN6 w - - bm Na3; hmvc 0; id "Predrag=N, JA._(=1031.57d3a1), SM._18.p WCCT7 Ty (tt), 8";
8/1p6/1Pp2N1q/p1Ppk2p/P3p3/3PPpPp/3K1P1P/2R5 w - - bm Ng4+; hmvc 0; id "Hasek, 1951_Emil Sutovsky, (after Hasek)_Minski CSS Michael Sch. RF, 9";
N7/8/2KQ2rp/6k1/3p3p/2p4P/4PP2/5N2 w - - bm f4+; hmvc 0; id "Nightmare II-28 - Endspiel D&_Krikheli, Jossif_Studie 1972, Anton CCC Festung, 10";
7k/1p5p/2qpn1PP/Pp1p4/8/3B2p1/3B1pP1/R4K2 w - - bm a6; hmvc 0; id "Nightmare II-09 - Patt_?_Studie?, 11";
2r2k2/8/8/3Rp1N1/2P3K1/8/2p4B/8 w - - bm Ne6+; hmvc 0; id "Nightmare II-16 - Endspiel T&_?_SF+tbs, 12";
1b6/2p5/1kB5/1PN5/8/3K4/8/q7 w - - bm Nb3; hmvc 0; id "Lejeune, Vincent_Hard, 2016_Nr.1 Festung, 13";
5N2/N1p2r2/5P2/Kp6/p7/8/2k1P3/2B5 w - - bm Ba3; hmvc 0; id "Akobi, Price?_Jon, Cooked?_?, 14";
2b1r3/r2ppN2/8/1p1p1k2/pP1P4/2P3R1/PP3PP1/2K5 w - - bm Nd6+; hmvc 0; id "ERET 093_Nightmare210, Festung_Simkhovich 1923, 15";
8/p7/kpP5/qrp1b3/rpP2b2/pP4b1/P3K3/8 w - - bm Kf3; hmvc 0; id "Sir, Roger_Penrose, Puzzle_?, 16";
3kn3/4pp2/n7/5B2/3N4/b6N/4K3/8 w - - bm Ng5; hmvc 0; id "Herbstmann, A.._Lejeune, V.._1956 2.price, CCC -2plies, 17";
8/1B3b1p/7P/p2p4/P2pp1kp/Q7/1N1Pq2P/7K w - - bm Qh3+; hmvc 0; id "Neghina, Mihai_P.K., 50J.T_Fishing Rod, 18";
3B4/1r2p3/r2p1p2/bkp1P1p1/1p1P1PPp/p1P4P/PPB1K3/8 w - - bm Ba4+; hmvc 0; id "Rudolph, Williams E._StingMcBrain, CritterChiron_1912, 19";
2b5/1pr4p/3bp1pk/1p6/1PpN2PP/K1P1n3/P3N1R1/3R4 w - - bm Nxb5; hmvc 0; id "Mihai Neghina_Desperation_White to move and draw, 20";
6k1/1qr1p2p/ppN5/3p1N2/6p1/1P1P2Pp/PP2PP1P/2RK4 w - - bm Ncxe7+; hmvc 0; id "Mihai Neghina, ACTA_Castle, 146 Stronghold_White to move and draw, 21";
8/b7/3p2b1/5p1n/5Pk1/4P1p1/B5P1/R4K2 w - - bm Ke2; hmvc 0; id "Mihai_Neghina_Festung, 22";

Here's how Stockfish 19 deals with these, single threaded (4.3GHz) in a single run of 30"/pos.:

Code: Select all

Bisher gelöst: 12 von 22  ;  6:20m

         1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------------
   0 |  10   7   -  16  17  20   -   0   0   -   -   -   -   0   0   0   -   -   0   4
  20 |   -   -

  K/s: 2.184.632  
  TotTime: 6:32m    SolTime: 6:20m

P.S. Had to log in 3 times again for submitting once

[chrisw]

Indeed the fortresses.epd there seems to be a rather inhomogeneous one, best moves missing to a big part too.

Problem with fortresses even more than with other tactical hard positions is, the margin between too easy and too hard for engines is even smaller. In big databases like HHdb eg., because these being composed for human players, these are to >99% too easy for modern engines, otherwise you could just take almost all the studies with claim "...to move and draw", getting fortresses that way plenty of at once. On the other hand, really (for engines) difficult ones to a big part are still too hard for STC engine- testing without Forward- Backward at all.
Here I have 22 examples of not much too easy and not much too hard ones filtered out of one mixed database of mine with first move only and some mimimum of sources given, checked for not having duplicates to fortresses.epd with epdutil:

Code: Select all

1N4K1/1pp1p3/7k/2r3pP/p2pp1P1/P6N/8/2B5 w - - bm Nc6; hmvc 0; id "Mihai Neghina_Standstill_Dort, 1";
8/8/3kp1pp/p2p1p2/Pp1P1P1P/1K4P1/8/8 w - - bm Kc2; hmvc 0; id "fort, ress_?_CCC, 2";
1k6/3p4/1B6/4Pp1p/1p5R/1p4p1/pP3n2/K6n w - - bm Rxh1; hmvc 0; id "eighth, step_to, heaven_?, 3";
K2b2r1/8/3p2p1/1Nk5/4Q3/p2P4/Pp2R1p1/2r2b2 w - - bm Kb7; hmvc 0; id "Krug, Peter_Stadlhofer, Anton_CSS, 4";
8/1p6/1Pp2N1q/p1Ppk2p/P3p3/3PPpPp/3K1P1P/1R6 w - - bm Ng4+; hmvc 0; id "Shredder, Chiron_Critter, nicht_Hasek 1951, 5";
4B3/k1K1pp2/p3bq2/2p1p3/1p2P1P1/4P2Q/nR6/8 w - - bm g5; hmvc 0; id "Neghina, Mihai_CSS, Forum_Selbstpatt, 6";
7B/7P/8/2p5/p1P2N2/P2p2pp/p1PPpnp1/2K1kbrq w - - bm Ba1; hmvc 0; id "Ziak=F_(=3344.58c1e1)_Schach Aktiv, Gull, 7";
7Q/4p3/4p3/p1p1P3/Pp2P3/3Kp3/p1PbP3/kN6 w - - bm Na3; hmvc 0; id "Predrag=N, JA._(=1031.57d3a1), SM._18.p WCCT7 Ty (tt), 8";
8/1p6/1Pp2N1q/p1Ppk2p/P3p3/3PPpPp/3K1P1P/2R5 w - - bm Ng4+; hmvc 0; id "Hasek, 1951_Emil Sutovsky, (after Hasek)_Minski CSS Michael Sch. RF, 9";
N7/8/2KQ2rp/6k1/3p3p/2p4P/4PP2/5N2 w - - bm f4+; hmvc 0; id "Nightmare II-28 - Endspiel D&_Krikheli, Jossif_Studie 1972, Anton CCC Festung, 10";
7k/1p5p/2qpn1PP/Pp1p4/8/3B2p1/3B1pP1/R4K2 w - - bm a6; hmvc 0; id "Nightmare II-09 - Patt_?_Studie?, 11";
2r2k2/8/8/3Rp1N1/2P3K1/8/2p4B/8 w - - bm Ne6+; hmvc 0; id "Nightmare II-16 - Endspiel T&_?_SF+tbs, 12";
1b6/2p5/1kB5/1PN5/8/3K4/8/q7 w - - bm Nb3; hmvc 0; id "Lejeune, Vincent_Hard, 2016_Nr.1 Festung, 13";
5N2/N1p2r2/5P2/Kp6/p7/8/2k1P3/2B5 w - - bm Ba3; hmvc 0; id "Akobi, Price?_Jon, Cooked?_?, 14";
2b1r3/r2ppN2/8/1p1p1k2/pP1P4/2P3R1/PP3PP1/2K5 w - - bm Nd6+; hmvc 0; id "ERET 093_Nightmare210, Festung_Simkhovich 1923, 15";
8/p7/kpP5/qrp1b3/rpP2b2/pP4b1/P3K3/8 w - - bm Kf3; hmvc 0; id "Sir, Roger_Penrose, Puzzle_?, 16";
3kn3/4pp2/n7/5B2/3N4/b6N/4K3/8 w - - bm Ng5; hmvc 0; id "Herbstmann, A.._Lejeune, V.._1956 2.price, CCC -2plies, 17";
8/1B3b1p/7P/p2p4/P2pp1kp/Q7/1N1Pq2P/7K w - - bm Qh3+; hmvc 0; id "Neghina, Mihai_P.K., 50J.T_Fishing Rod, 18";
3B4/1r2p3/r2p1p2/bkp1P1p1/1p1P1PPp/p1P4P/PPB1K3/8 w - - bm Ba4+; hmvc 0; id "Rudolph, Williams E._StingMcBrain, CritterChiron_1912, 19";
2b5/1pr4p/3bp1pk/1p6/1PpN2PP/K1P1n3/P3N1R1/3R4 w - - bm Nxb5; hmvc 0; id "Mihai Neghina_Desperation_White to move and draw, 20";
6k1/1qr1p2p/ppN5/3p1N2/6p1/1P1P2Pp/PP2PP1P/2RK4 w - - bm Ncxe7+; hmvc 0; id "Mihai Neghina, ACTA_Castle, 146 Stronghold_White to move and draw, 21";
8/b7/3p2b1/5p1n/5Pk1/4P1p1/B5P1/R4K2 w - - bm Ke2; hmvc 0; id "Mihai_Neghina_Festung, 22";

Here's how Stockfish 19 deals with these, single threaded (4.3GHz) in a single run of 30"/pos.:

Code: Select all

Bisher gelöst: 12 von 22  ;  6:20m

         1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------------
   0 |  10   7   -  16  17  20   -   0   0   -   -   -   -   0   0   0   -   -   0   4
  20 |   -   -

  K/s: 2.184.632  
  TotTime: 6:32m    SolTime: 6:20m

P.S. Had to log in 3 times again for submitting once

Thanks, added to the github - epds, with a credit

[peter]

47 more of that kind:

Code: Select all

7k/4PBNp/7P/2p5/2P5/2rq2Pp/1P3P1K/8 b - - bm Qxg3+; hmvc 0; id "Sting, Alone_Neghina's, Chase_derived, VL CCC, 1";
2r2r1k/qb3p1p/p2p1Np1/np1Pp1P1/2P5/3RP2P/5R2/4Q1K1 w - - bm Rf4; hmvc 0; id "Rybka, Forum_Corbit, Dann_CCC recent positions 6, 2";
3r3r/k7/2p5/1pPp1p1p/pP2pPpP/P1K1P1P1/2b5/BN6 w - - bm Kd4; hmvc 0; id "Rybka, Forum_Corbit, Dann_Recent 7, 3";
N3q1kn/1N1p1p2/3PpPp1/Bp2P1P1/b7/P5p1/1P4Pb/7K w - - bm Nd8; hmvc 0; id "Neghina, Mihai_Winterferien, rätsel1_White to move and draw, 4";
8/1p4p1/1Np2pP1/1p6/6P1/P2kP1PK/P2P4/2b5 w - - bm Na8; hmvc 0; id "Gelly, M.._Eigenmann, W.._CSS LC0+Eret, 5";
2br4/r2pp3/8/1p1p1kN1/pP1P4/2P3R1/PP3PP1/2K5 w - - bm Nf7; hmvc 0; id "Simkovich_Eret93, Minus1Zug_1923, 6";
1b6/r2p1k2/8/1R1b3R/8/1n1K4/8/1nB5 w - - bm Bb2; hmvc 0; id "ACT, 94._??, ??._?, 8";
2kbK2R/4p3/1p2Pr2/1p1P4/1PN5/1r6/5P2/8 w - - bm Nd6+; hmvc 0; id "ACT, 113_??, ??._?, 9";
2n3R1/5p2/P7/5P1k/4BK1p/2q3Pp/7P/8 w - - bm Rh8+; hmvc 0; id "ACT, 114_?_?, 10";
6rq/6p1/pQp1p1p1/P1p1P2p/p1Pk3B/N7/3K4/1b5n w - - bm Nc2+; hmvc 0; id "ACT, 150_??, ??._?, 11";
8/8/8/8/2K1N3/5Bpp/3N1p2/4kr1n w - - bm Kc3; hmvc 0; id "ACT, 180_?_Alternative Chess Test-Analysis, 12";
k4N2/1p3b2/p7/P1K1P3/p1P2N2/8/1b6/8 w - - bm e6; hmvc 0; id "ACT, 203_??, ??._Tbs? Drawing Eval!, 13";
k7/8/KP6/8/5b2/2R5/1rNP4/5r1N w - - bm Nf2; hmvc 0; id "ACT, 205_??, ??._Tbs, Drawing Eval, 14";
rb2n3/p3B3/1P5p/2P4K/5k2/1r4p1/3R2R1/8 w - - bm Rxg3; hmvc 0; id "Timman, J.._HHdbVI, ??._?, 15";
8/7P/1P5B/2B1Q1n1/3nn2P/1PRnk1nR/3nnnK1/2B1nQBn w - - bm Qxe1; hmvc 0; id "Krabbé, Tim_Patt, Matt_Roland Riener CSS, 16";
3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - bm Ng6+; hmvc 0; id "Tough, Tactical_Test, Toughest_CCC, 17";
5N1n/3P1p1p/5P1k/1Pp1P2p/4KP1P/2p5/p4b2/5N2 w - - bm Ne3; hmvc 0; id "Michal Formela_?_Polski Zwiazek Szachowy 2019 (nr 1856,, 18";
k7/7p/2p5/2rp1PP1/pp6/7p/5K1P/4R3 w - - bm g6; hmvc 0; id "Zepler, Erich Ernest_1928_?, 19";
1r6/1n1R1b2/8/1p1p3k/pPpPp1p1/2P1P3/P2K1PP1/8 w - - bm Rxb7; hmvc 0; id "Illescas, Miguel_Eigenmann, Walter_Studie 2018 CSS, E.C.2-18, 20";
n6r/b7/1R4R1/2k5/2pppppp/8/1PPPPPP1/r4NK1 w - - bm b4+; hmvc 0; id "Korolkov_Riener_?, 21";
r1n5/k3pp2/2Rr4/1p6/1p6/p4pp1/QP4Pn/4RBK1 w - - bm Rec1; hmvc 0; id "Krug, P.._Engine, Crackers2_Nr.20, Studie 2018, 22";
8/1pp1p3/6p1/3P4/pP3P2/8/k1K3P1/8 b - - bm c6; hmvc 0; id "Grigoriev, N.._EC2, Nr.3_N. Grigoriev 1792 bm+am, 23";
8/6p1/5p2/k5p1/P7/2K2r1P/3R1P2/8 w - - bm Kc2; hmvc 0; id "J, P.._How, long?_CCC, Crystal, 24";
3k4/3p4/3p1K2/NN1p4/8/8/4p3/8 w - - bm Nb7+; hmvc 0; id "Matisons, Hermanis_Studie, YACPDB_ChessPlayer CSS, 25";
3n1k2/p2B4/Pp1P2pp/1P1ppp2/7P/4P1P1/5PK1/8 w - - bm f4; hmvc 0; id "Berger, Peter_Single, Best_+Avoid ...Bc6?, 26";
rr4k1/3nb1pp/4pn2/2ppP1B1/pP6/P1PP4/5PPP/R3RNK1 b - - bm Nxe5; hmvc 0; id "Combusken, Will_Counter, BlackWin?_Franks Turnier, WE CSS, 27";
r7/1b1r4/k1p1p1p1/1p1pPpPp/p1PP1P1P/PP1K4/8/4Q3 w - - bm Qa5+; hmvc 0; id "IQ, 45._Geänderte, Stellung?_?, 28";
r7/8/8/5bk1/8/5B2/5RPP/6K1 b - - bm Ra1+; hmvc 0; id "6, Steiner_puzzle, ?._Chessqueen CCC, 29";
8/p5p1/Pp1k1n2/1Pp2BP1/2Pp1K2/3P4/8/8 w - - bm gxf6; hmvc 0; id "David, Anton_Dimitrij, Kollars_?, 30";
8/2pp1ppp/8/1pPP1PPk/8/8/PP6/6K1 w - - bm g6; hmvc 0; id "Troitzky, A.._Cooked, Acase_CCC, 31";
nq6/pPpB4/2K5/1P2BN2/3Pk2p/5p2/8/6bN w - - bm b6; hmvc 0; id "Acase, CCC_After 10, moves of_Yuri Bazlow 2019, 32";
R3K1Bq/pPp1B1N1/8/1P5p/1rnP4/4kp2/8/6bN w - - bm Nf5+; am b8=Q; hmvc 0; id "Yuri Bazlov, Russia_?_1st Prize Mikitovics 70 JT 2019 sect.dr, 33";
B7/4p3/p1Pp2Pp/3pP2k/4r3/Kp6/1PbP1P2/8 w - - bm d3; hmvc 0; id "PervakovO, WhiteM+W_Acase, Cooked_1.sp.p Averbach-80 JT EBUR, 34";
r5b1/7P/p7/2B4p/rp1R4/4N1P1/pk6/3K4 w - - bm Rd2+; hmvc 0; id "Zemljanski, Y.I._Originalstellung, Acase h6_Tsjechover-90 MT Shakhmaty Peterbu, 35";
3B4/1r2p3/r2p1p2/bkp1P1p1/1p1P1PPp/p1P1K2P/PPB5/8 w - - bm Ba4+; hmvc 0; id "Rudolph, W.E._?_Studie 1912 HTC108-30 S, 36";
1q1k4/1n1p1b2/p1pPp1p1/r1P1PpPp/p1P2P1N/P1QK4/8/4B3 w - - bm Qxa5+; hmvc 0; id "Original, Partie?_Festung, Suite?_ACT4 69, 37";
1q6/2R4K/1k1p1p2/2NPpP1n/1p1pBp2/1PbP1P2/8/B7 w - - bm Rb7+; hmvc 0; id "Hobacle, White_ToMove, AndDraw_?, 38";
3r4/6bk/R6p/3Bp2P/4P1N1/p5P1/5PK1/r7 b - - bm Re8; hmvc 0; id "Ljubicic, L.._Blass, U.._ICCF 2014 corr. WM, 39";
7k/Bp1R1b1p/1p2pK1n/8/4P1P1/6pB/4p3/8 w - - bm Bb8; hmvc 0; id "Eigenmann, W._NN, Ethereal14_CSS, 40";
rkb5/1p6/pBp5/PPPp4/1K1Pp1P1/4Pp1p/5P1P/8 w - - bm Ba7+; hmvc 0; id "Hasek=J_(=0340.87b4b8)_2.p L'Italia Scacchistica#287, 41";
6bk/5pp1/3N1p1b/8/4p1p1/5PPp/4pPNP/4K3 w - - bm Nh4; hmvc 0; id "Koranyi=A_(=0062.47e1h8)_.c Arguelles JT Problemas (SEPA), 42";
8/p1p3p1/1p1kpp2/6Pp/3N1P2/5K2/7P/8 w - - bm g6; hmvc 0; id "Formela=M_(=0001.37f3d6) U2_Sinfonie Scacchistiche#1097, 43";
3N1B2/p2p4/n4p2/pK2p3/3kp3/1P2RPp1/3PP2p/b7 w - - bm Rd3+; hmvc 0; id "Popov=A_(=0144.48b5d4)_Zadachy i Etyudy, 44";
8/3p2p1/2pp2P1/1p1p2P1/6KP/p5P1/4B1P1/1k6 w - - bm Bc4; hmvc 0; id "Gatti=D_(=0010.57g4b1)_Caputto-95 JT 7th UAPA internet ty#20, 45";
r1rbk1K1/p7/Pp1P2R1/1b4N1/5Bp1/2PR2P1/2P4p/8 w - - bm Rg7; hmvc 0; id "Timman=J_(=0871.54g8e8)_1.p Problem Paradise#081, 46";
6Q1/PK1kn3/p1p2b2/1bb1b3/6p1/8/5b1b/8 w - - bm a8=B; hmvc 0; id "Gatti=D_(=1093.13b7d7)_Best Problems, 47";
8/2pB2p1/p1np2Pb/1p1P1Rp1/2p2p1K/5k1P/2PP1P2/1b1R4 w - - bm Kh5; hmvc 0; id "Gelly=M_(=0273.68h4f3)_Polish Chess Federation ty#0556, 48";

And again a sinlge run of Stockfish 19 single threaded with 30"/pos:

Code: Select all

Bisher gelöst: 28 von 47 

         1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------------
   0 |   1   0   0   0   -   0   0  16   0   -   -   0   0   2   -   -   -   -   -   -
  20 |   -   8   -   0   3   0   2   0   0   0   -   0   -   0   3   0   0   -   -   0
  40 |   -   1   -   2   9   -   -  

0 doublets to older .epd on github- site again, regards

[chrisw]

47 more of that kind:

Code: Select all

7k/4PBNp/7P/2p5/2P5/2rq2Pp/1P3P1K/8 b - - bm Qxg3+; hmvc 0; id "Sting, Alone_Neghina's, Chase_derived, VL CCC, 1";
2r2r1k/qb3p1p/p2p1Np1/np1Pp1P1/2P5/3RP2P/5R2/4Q1K1 w - - bm Rf4; hmvc 0; id "Rybka, Forum_Corbit, Dann_CCC recent positions 6, 2";
3r3r/k7/2p5/1pPp1p1p/pP2pPpP/P1K1P1P1/2b5/BN6 w - - bm Kd4; hmvc 0; id "Rybka, Forum_Corbit, Dann_Recent 7, 3";
N3q1kn/1N1p1p2/3PpPp1/Bp2P1P1/b7/P5p1/1P4Pb/7K w - - bm Nd8; hmvc 0; id "Neghina, Mihai_Winterferien, rätsel1_White to move and draw, 4";
8/1p4p1/1Np2pP1/1p6/6P1/P2kP1PK/P2P4/2b5 w - - bm Na8; hmvc 0; id "Gelly, M.._Eigenmann, W.._CSS LC0+Eret, 5";
2br4/r2pp3/8/1p1p1kN1/pP1P4/2P3R1/PP3PP1/2K5 w - - bm Nf7; hmvc 0; id "Simkovich_Eret93, Minus1Zug_1923, 6";
1b6/r2p1k2/8/1R1b3R/8/1n1K4/8/1nB5 w - - bm Bb2; hmvc 0; id "ACT, 94._??, ??._?, 8";
2kbK2R/4p3/1p2Pr2/1p1P4/1PN5/1r6/5P2/8 w - - bm Nd6+; hmvc 0; id "ACT, 113_??, ??._?, 9";
2n3R1/5p2/P7/5P1k/4BK1p/2q3Pp/7P/8 w - - bm Rh8+; hmvc 0; id "ACT, 114_?_?, 10";
6rq/6p1/pQp1p1p1/P1p1P2p/p1Pk3B/N7/3K4/1b5n w - - bm Nc2+; hmvc 0; id "ACT, 150_??, ??._?, 11";
8/8/8/8/2K1N3/5Bpp/3N1p2/4kr1n w - - bm Kc3; hmvc 0; id "ACT, 180_?_Alternative Chess Test-Analysis, 12";
k4N2/1p3b2/p7/P1K1P3/p1P2N2/8/1b6/8 w - - bm e6; hmvc 0; id "ACT, 203_??, ??._Tbs? Drawing Eval!, 13";
k7/8/KP6/8/5b2/2R5/1rNP4/5r1N w - - bm Nf2; hmvc 0; id "ACT, 205_??, ??._Tbs, Drawing Eval, 14";
rb2n3/p3B3/1P5p/2P4K/5k2/1r4p1/3R2R1/8 w - - bm Rxg3; hmvc 0; id "Timman, J.._HHdbVI, ??._?, 15";
8/7P/1P5B/2B1Q1n1/3nn2P/1PRnk1nR/3nnnK1/2B1nQBn w - - bm Qxe1; hmvc 0; id "Krabbé, Tim_Patt, Matt_Roland Riener CSS, 16";
3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - bm Ng6+; hmvc 0; id "Tough, Tactical_Test, Toughest_CCC, 17";
5N1n/3P1p1p/5P1k/1Pp1P2p/4KP1P/2p5/p4b2/5N2 w - - bm Ne3; hmvc 0; id "Michal Formela_?_Polski Zwiazek Szachowy 2019 (nr 1856,, 18";
k7/7p/2p5/2rp1PP1/pp6/7p/5K1P/4R3 w - - bm g6; hmvc 0; id "Zepler, Erich Ernest_1928_?, 19";
1r6/1n1R1b2/8/1p1p3k/pPpPp1p1/2P1P3/P2K1PP1/8 w - - bm Rxb7; hmvc 0; id "Illescas, Miguel_Eigenmann, Walter_Studie 2018 CSS, E.C.2-18, 20";
n6r/b7/1R4R1/2k5/2pppppp/8/1PPPPPP1/r4NK1 w - - bm b4+; hmvc 0; id "Korolkov_Riener_?, 21";
r1n5/k3pp2/2Rr4/1p6/1p6/p4pp1/QP4Pn/4RBK1 w - - bm Rec1; hmvc 0; id "Krug, P.._Engine, Crackers2_Nr.20, Studie 2018, 22";
8/1pp1p3/6p1/3P4/pP3P2/8/k1K3P1/8 b - - bm c6; hmvc 0; id "Grigoriev, N.._EC2, Nr.3_N. Grigoriev 1792 bm+am, 23";
8/6p1/5p2/k5p1/P7/2K2r1P/3R1P2/8 w - - bm Kc2; hmvc 0; id "J, P.._How, long?_CCC, Crystal, 24";
3k4/3p4/3p1K2/NN1p4/8/8/4p3/8 w - - bm Nb7+; hmvc 0; id "Matisons, Hermanis_Studie, YACPDB_ChessPlayer CSS, 25";
3n1k2/p2B4/Pp1P2pp/1P1ppp2/7P/4P1P1/5PK1/8 w - - bm f4; hmvc 0; id "Berger, Peter_Single, Best_+Avoid ...Bc6?, 26";
rr4k1/3nb1pp/4pn2/2ppP1B1/pP6/P1PP4/5PPP/R3RNK1 b - - bm Nxe5; hmvc 0; id "Combusken, Will_Counter, BlackWin?_Franks Turnier, WE CSS, 27";
r7/1b1r4/k1p1p1p1/1p1pPpPp/p1PP1P1P/PP1K4/8/4Q3 w - - bm Qa5+; hmvc 0; id "IQ, 45._Geänderte, Stellung?_?, 28";
r7/8/8/5bk1/8/5B2/5RPP/6K1 b - - bm Ra1+; hmvc 0; id "6, Steiner_puzzle, ?._Chessqueen CCC, 29";
8/p5p1/Pp1k1n2/1Pp2BP1/2Pp1K2/3P4/8/8 w - - bm gxf6; hmvc 0; id "David, Anton_Dimitrij, Kollars_?, 30";
8/2pp1ppp/8/1pPP1PPk/8/8/PP6/6K1 w - - bm g6; hmvc 0; id "Troitzky, A.._Cooked, Acase_CCC, 31";
nq6/pPpB4/2K5/1P2BN2/3Pk2p/5p2/8/6bN w - - bm b6; hmvc 0; id "Acase, CCC_After 10, moves of_Yuri Bazlow 2019, 32";
R3K1Bq/pPp1B1N1/8/1P5p/1rnP4/4kp2/8/6bN w - - bm Nf5+; am b8=Q; hmvc 0; id "Yuri Bazlov, Russia_?_1st Prize Mikitovics 70 JT 2019 sect.dr, 33";
B7/4p3/p1Pp2Pp/3pP2k/4r3/Kp6/1PbP1P2/8 w - - bm d3; hmvc 0; id "PervakovO, WhiteM+W_Acase, Cooked_1.sp.p Averbach-80 JT EBUR, 34";
r5b1/7P/p7/2B4p/rp1R4/4N1P1/pk6/3K4 w - - bm Rd2+; hmvc 0; id "Zemljanski, Y.I._Originalstellung, Acase h6_Tsjechover-90 MT Shakhmaty Peterbu, 35";
3B4/1r2p3/r2p1p2/bkp1P1p1/1p1P1PPp/p1P1K2P/PPB5/8 w - - bm Ba4+; hmvc 0; id "Rudolph, W.E._?_Studie 1912 HTC108-30 S, 36";
1q1k4/1n1p1b2/p1pPp1p1/r1P1PpPp/p1P2P1N/P1QK4/8/4B3 w - - bm Qxa5+; hmvc 0; id "Original, Partie?_Festung, Suite?_ACT4 69, 37";
1q6/2R4K/1k1p1p2/2NPpP1n/1p1pBp2/1PbP1P2/8/B7 w - - bm Rb7+; hmvc 0; id "Hobacle, White_ToMove, AndDraw_?, 38";
3r4/6bk/R6p/3Bp2P/4P1N1/p5P1/5PK1/r7 b - - bm Re8; hmvc 0; id "Ljubicic, L.._Blass, U.._ICCF 2014 corr. WM, 39";
7k/Bp1R1b1p/1p2pK1n/8/4P1P1/6pB/4p3/8 w - - bm Bb8; hmvc 0; id "Eigenmann, W._NN, Ethereal14_CSS, 40";
rkb5/1p6/pBp5/PPPp4/1K1Pp1P1/4Pp1p/5P1P/8 w - - bm Ba7+; hmvc 0; id "Hasek=J_(=0340.87b4b8)_2.p L'Italia Scacchistica#287, 41";
6bk/5pp1/3N1p1b/8/4p1p1/5PPp/4pPNP/4K3 w - - bm Nh4; hmvc 0; id "Koranyi=A_(=0062.47e1h8)_.c Arguelles JT Problemas (SEPA), 42";
8/p1p3p1/1p1kpp2/6Pp/3N1P2/5K2/7P/8 w - - bm g6; hmvc 0; id "Formela=M_(=0001.37f3d6) U2_Sinfonie Scacchistiche#1097, 43";
3N1B2/p2p4/n4p2/pK2p3/3kp3/1P2RPp1/3PP2p/b7 w - - bm Rd3+; hmvc 0; id "Popov=A_(=0144.48b5d4)_Zadachy i Etyudy, 44";
8/3p2p1/2pp2P1/1p1p2P1/6KP/p5P1/4B1P1/1k6 w - - bm Bc4; hmvc 0; id "Gatti=D_(=0010.57g4b1)_Caputto-95 JT 7th UAPA internet ty#20, 45";
r1rbk1K1/p7/Pp1P2R1/1b4N1/5Bp1/2PR2P1/2P4p/8 w - - bm Rg7; hmvc 0; id "Timman=J_(=0871.54g8e8)_1.p Problem Paradise#081, 46";
6Q1/PK1kn3/p1p2b2/1bb1b3/6p1/8/5b1b/8 w - - bm a8=B; hmvc 0; id "Gatti=D_(=1093.13b7d7)_Best Problems, 47";
8/2pB2p1/p1np2Pb/1p1P1Rp1/2p2p1K/5k1P/2PP1P2/1b1R4 w - - bm Kh5; hmvc 0; id "Gelly=M_(=0273.68h4f3)_Polish Chess Federation ty#0556, 48";

And again a sinlge run of Stockfish 19 single threaded with 30"/pos:

Code: Select all

Bisher gelöst: 28 von 47 

         1   2   3   4   5   6   7   8   9  10  11  12  13  14  15  16  17  18  19  20
 -------------------------------------------------------------------------------------
   0 |   1   0   0   0   -   0   0  16   0   -   -   0   0   2   -   -   -   -   -   -
  20 |   -   8   -   0   3   0   2   0   0   0   -   0   -   0   3   0   0   -   -   0
  40 |   -   1   -   2   9   -   -  

0 doublets to older .epd on github- site again, regards

Great, added

[Ajedrecista]

Hello Peter:

47 more of that kind:

Code: Select all

7k/4PBNp/7P/2p5/2P5/2rq2Pp/1P3P1K/8 b - - bm Qxg3+; hmvc 0; id "Sting, Alone_Neghina's, Chase_derived, VL CCC, 1";
2r2r1k/qb3p1p/p2p1Np1/np1Pp1P1/2P5/3RP2P/5R2/4Q1K1 w - - bm Rf4; hmvc 0; id "Rybka, Forum_Corbit, Dann_CCC recent positions 6, 2";
3r3r/k7/2p5/1pPp1p1p/pP2pPpP/P1K1P1P1/2b5/BN6 w - - bm Kd4; hmvc 0; id "Rybka, Forum_Corbit, Dann_Recent 7, 3";
N3q1kn/1N1p1p2/3PpPp1/Bp2P1P1/b7/P5p1/1P4Pb/7K w - - bm Nd8; hmvc 0; id "Neghina, Mihai_Winterferien, rätsel1_White to move and draw, 4";
8/1p4p1/1Np2pP1/1p6/6P1/P2kP1PK/P2P4/2b5 w - - bm Na8; hmvc 0; id "Gelly, M.._Eigenmann, W.._CSS LC0+Eret, 5";
2br4/r2pp3/8/1p1p1kN1/pP1P4/2P3R1/PP3PP1/2K5 w - - bm Nf7; hmvc 0; id "Simkovich_Eret93, Minus1Zug_1923, 6";
1b6/r2p1k2/8/1R1b3R/8/1n1K4/8/1nB5 w - - bm Bb2; hmvc 0; id "ACT, 94._??, ??._?, 8";
2kbK2R/4p3/1p2Pr2/1p1P4/1PN5/1r6/5P2/8 w - - bm Nd6+; hmvc 0; id "ACT, 113_??, ??._?, 9";
2n3R1/5p2/P7/5P1k/4BK1p/2q3Pp/7P/8 w - - bm Rh8+; hmvc 0; id "ACT, 114_?_?, 10";
6rq/6p1/pQp1p1p1/P1p1P2p/p1Pk3B/N7/3K4/1b5n w - - bm Nc2+; hmvc 0; id "ACT, 150_??, ??._?, 11";
8/8/8/8/2K1N3/5Bpp/3N1p2/4kr1n w - - bm Kc3; hmvc 0; id "ACT, 180_?_Alternative Chess Test-Analysis, 12";
k4N2/1p3b2/p7/P1K1P3/p1P2N2/8/1b6/8 w - - bm e6; hmvc 0; id "ACT, 203_??, ??._Tbs? Drawing Eval!, 13";
k7/8/KP6/8/5b2/2R5/1rNP4/5r1N w - - bm Nf2; hmvc 0; id "ACT, 205_??, ??._Tbs, Drawing Eval, 14";
rb2n3/p3B3/1P5p/2P4K/5k2/1r4p1/3R2R1/8 w - - bm Rxg3; hmvc 0; id "Timman, J.._HHdbVI, ??._?, 15";
8/7P/1P5B/2B1Q1n1/3nn2P/1PRnk1nR/3nnnK1/2B1nQBn w - - bm Qxe1; hmvc 0; id "Krabbé, Tim_Patt, Matt_Roland Riener CSS, 16";
3rr1bk/6np/2p1n3/1pPpNpPp/pP1QpP1P/P2PP3/2K5/B7 w - - bm Ng6+; hmvc 0; id "Tough, Tactical_Test, Toughest_CCC, 17";
5N1n/3P1p1p/5P1k/1Pp1P2p/4KP1P/2p5/p4b2/5N2 w - - bm Ne3; hmvc 0; id "Michal Formela_?_Polski Zwiazek Szachowy 2019 (nr 1856,, 18";
k7/7p/2p5/2rp1PP1/pp6/7p/5K1P/4R3 w - - bm g6; hmvc 0; id "Zepler, Erich Ernest_1928_?, 19";
1r6/1n1R1b2/8/1p1p3k/pPpPp1p1/2P1P3/P2K1PP1/8 w - - bm Rxb7; hmvc 0; id "Illescas, Miguel_Eigenmann, Walter_Studie 2018 CSS, E.C.2-18, 20";
n6r/b7/1R4R1/2k5/2pppppp/8/1PPPPPP1/r4NK1 w - - bm b4+; hmvc 0; id "Korolkov_Riener_?, 21";
r1n5/k3pp2/2Rr4/1p6/1p6/p4pp1/QP4Pn/4RBK1 w - - bm Rec1; hmvc 0; id "Krug, P.._Engine, Crackers2_Nr.20, Studie 2018, 22";
8/1pp1p3/6p1/3P4/pP3P2/8/k1K3P1/8 b - - bm c6; hmvc 0; id "Grigoriev, N.._EC2, Nr.3_N. Grigoriev 1792 bm+am, 23";
8/6p1/5p2/k5p1/P7/2K2r1P/3R1P2/8 w - - bm Kc2; hmvc 0; id "J, P.._How, long?_CCC, Crystal, 24";
3k4/3p4/3p1K2/NN1p4/8/8/4p3/8 w - - bm Nb7+; hmvc 0; id "Matisons, Hermanis_Studie, YACPDB_ChessPlayer CSS, 25";
3n1k2/p2B4/Pp1P2pp/1P1ppp2/7P/4P1P1/5PK1/8 w - - bm f4; hmvc 0; id "Berger, Peter_Single, Best_+Avoid ...Bc6?, 26";
rr4k1/3nb1pp/4pn2/2ppP1B1/pP6/P1PP4/5PPP/R3RNK1 b - - bm Nxe5; hmvc 0; id "Combusken, Will_Counter, BlackWin?_Franks Turnier, WE CSS, 27";
r7/1b1r4/k1p1p1p1/1p1pPpPp/p1PP1P1P/PP1K4/8/4Q3 w - - bm Qa5+; hmvc 0; id "IQ, 45._Geänderte, Stellung?_?, 28";
r7/8/8/5bk1/8/5B2/5RPP/6K1 b - - bm Ra1+; hmvc 0; id "6, Steiner_puzzle, ?._Chessqueen CCC, 29";
8/p5p1/Pp1k1n2/1Pp2BP1/2Pp1K2/3P4/8/8 w - - bm gxf6; hmvc 0; id "David, Anton_Dimitrij, Kollars_?, 30";
8/2pp1ppp/8/1pPP1PPk/8/8/PP6/6K1 w - - bm g6; hmvc 0; id "Troitzky, A.._Cooked, Acase_CCC, 31";
nq6/pPpB4/2K5/1P2BN2/3Pk2p/5p2/8/6bN w - - bm b6; hmvc 0; id "Acase, CCC_After 10, moves of_Yuri Bazlow 2019, 32";
R3K1Bq/pPp1B1N1/8/1P5p/1rnP4/4kp2/8/6bN w - - bm Nf5+; am b8=Q; hmvc 0; id "Yuri Bazlov, Russia_?_1st Prize Mikitovics 70 JT 2019 sect.dr, 33";
B7/4p3/p1Pp2Pp/3pP2k/4r3/Kp6/1PbP1P2/8 w - - bm d3; hmvc 0; id "PervakovO, WhiteM+W_Acase, Cooked_1.sp.p Averbach-80 JT EBUR, 34";
r5b1/7P/p7/2B4p/rp1R4/4N1P1/pk6/3K4 w - - bm Rd2+; hmvc 0; id "Zemljanski, Y.I._Originalstellung, Acase h6_Tsjechover-90 MT Shakhmaty Peterbu, 35";
3B4/1r2p3/r2p1p2/bkp1P1p1/1p1P1PPp/p1P1K2P/PPB5/8 w - - bm Ba4+; hmvc 0; id "Rudolph, W.E._?_Studie 1912 HTC108-30 S, 36";
1q1k4/1n1p1b2/p1pPp1p1/r1P1PpPp/p1P2P1N/P1QK4/8/4B3 w - - bm Qxa5+; hmvc 0; id "Original, Partie?_Festung, Suite?_ACT4 69, 37";
1q6/2R4K/1k1p1p2/2NPpP1n/1p1pBp2/1PbP1P2/8/B7 w - - bm Rb7+; hmvc 0; id "Hobacle, White_ToMove, AndDraw_?, 38";
3r4/6bk/R6p/3Bp2P/4P1N1/p5P1/5PK1/r7 b - - bm Re8; hmvc 0; id "Ljubicic, L.._Blass, U.._ICCF 2014 corr. WM, 39";
7k/Bp1R1b1p/1p2pK1n/8/4P1P1/6pB/4p3/8 w - - bm Bb8; hmvc 0; id "Eigenmann, W._NN, Ethereal14_CSS, 40";
rkb5/1p6/pBp5/PPPp4/1K1Pp1P1/4Pp1p/5P1P/8 w - - bm Ba7+; hmvc 0; id "Hasek=J_(=0340.87b4b8)_2.p L'Italia Scacchistica#287, 41";
6bk/5pp1/3N1p1b/8/4p1p1/5PPp/4pPNP/4K3 w - - bm Nh4; hmvc 0; id "Koranyi=A_(=0062.47e1h8)_.c Arguelles JT Problemas (SEPA), 42";
8/p1p3p1/1p1kpp2/6Pp/3N1P2/5K2/7P/8 w - - bm g6; hmvc 0; id "Formela=M_(=0001.37f3d6) U2_Sinfonie Scacchistiche#1097, 43";
3N1B2/p2p4/n4p2/pK2p3/3kp3/1P2RPp1/3PP2p/b7 w - - bm Rd3+; hmvc 0; id "Popov=A_(=0144.48b5d4)_Zadachy i Etyudy, 44";
8/3p2p1/2pp2P1/1p1p2P1/6KP/p5P1/4B1P1/1k6 w - - bm Bc4; hmvc 0; id "Gatti=D_(=0010.57g4b1)_Caputto-95 JT 7th UAPA internet ty#20, 45";
r1rbk1K1/p7/Pp1P2R1/1b4N1/5Bp1/2PR2P1/2P4p/8 w - - bm Rg7; hmvc 0; id "Timman=J_(=0871.54g8e8)_1.p Problem Paradise#081, 46";
6Q1/PK1kn3/p1p2b2/1bb1b3/6p1/8/5b1b/8 w - - bm a8=B; hmvc 0; id "Gatti=D_(=1093.13b7d7)_Best Problems, 47";
8/2pB2p1/p1np2Pb/1p1P1Rp1/2p2p1K/5k1P/2PP1P2/1b1R4 w - - bm Kh5; hmvc 0; id "Gelly=M_(=0273.68h4f3)_Polish Chess Federation ty#0556, 48";

[...]

Regarding your last EPD, there are 47 positions numbered from 1 to 6 and from 8 to 48. Number 7 is missing and numbering afterwards is shifted by 1.

I looked to some ??, ?? sources and luckily found your number 9 (ACT, 113) to be YACPDB #285919 thanks to an Arves article. I did not find the original publication, but I found the reprint cited at YACPDB, which relates this problem with the number 6 of your EPD.

I also found your number 10 (ACT, 114) to be YACPDB #275949. Then, I found the Arves article where it is (author Kondratjev).

I also found your number 11 (ACT, 150) to be a correction from a problem of 1948 thanks to an Arves article.

I also found your number 13 (ACT, 203) to be YACPDB #329681 thanks to an Arves article. I did not find the original publication.

Regards from Spain.

Ajedrecista.

[peter]

Regarding your last EPD, there are 47 positions numbered from 1 to 6 and from 8 to 48. Number 7 is missing and numbering afterwards is shifted by 1.

I looked to some ??, ?? sources and luckily found...

I filtered the positions at first glance out of a mixed problem- database of mine, then exported first moves and headers by pgn2epd, then looked closer at the level of difficulty for engines and removed the ones too easy and the ones too hard for STC- engine- testing. That's where the discontinuing numbers come from, by exporting to .epd the ones before removing of some in between were kept.
And I didn't have much search for their original sources beyond the ones stored in mixed database already, just uploaded after quick check with 30"- SF- single thread- run. Thank's for adding your findings, regards

[chrisw]

Regarding your last EPD, there are 47 positions numbered from 1 to 6 and from 8 to 48. Number 7 is missing and numbering afterwards is shifted by 1.

I looked to some ??, ?? sources and luckily found...

I filtered the positions at first glance out of a mixed problem- database of mine, then exported first moves and headers by pgn2epd, then looked closer at the level of difficulty for engines and removed the ones too easy and the ones too hard for STC- engine- testing. That's where the discontinuing numbers come from, by exporting to .epd the ones before removing of some in between were kept.
And I didn't have much search for their original sources beyond the ones stored in mixed database already, just uploaded after quick check with 30"- SF- single thread- run. Thank's for adding your findings, regards

Can this process be automated? I guess if you need 30 seconds, that’s a throughput of 2 positions per second on a 64 core? Then the problem becomes one of the frequency of fortresses per N positions. Not very many presumably.

So, pre-filter the position suite. Let’s say we scan standard SF/Leela binpacks. Suppose we look for only 16 pieces left or fewer. If abs(V) > 100 cp (guessing), but result = 0 then that’s a candidate for your 30 second filter. Or am I being dumb here?

[peter]

Can this process be automated?

Guess, you didn't fully get the way of my findings being done by me here, let me tell more about it in detail so. The database of mine, which I started the search in as a first trial, is one of mixed problems, most of them with 1-0 and 0-1- results or lines without given decision for one side. First thing I filtered out was, if it's kind of single best move puzzles at all, which is the case with most of the positions within this one database, second, if it's for sure drawn as for outcomes, thus the probability, it's fortress- positions was already rather high. But not until then I have to judge for difficulty engine- wise and this one final process is the only one you can get some automatic judgement by engine. Not before I had filtered out the real fortresses manually out of the thematically already prefiltered database, that I have collected (manually) over the years, I could have started the engine- wise judgement. So the amount of then finally by engine tested positions and found difficult enough, is as small as you see with the 69 given, out of a database, I started with of overall 1634 positions big all in all, you see?

A much bigger collection of testable positions you'd get out of databases like HHdb, (103157 positions) most of the drawn ones will be fortresses too, but for human players composed, not for engines, >99% of these fortresses get solved by modern engines single threaded at once, so not to be used for engine- testing anyhow, as I wrote in my first posting in this thread already. Together with the few for engine- testing usable positions out of big databases of composed studies not too easy, you'll get some, that are still too difficult for STC- engine- testing, those simply aren't solved by engines still at all, if not with very much hardware- time and or forward- backward- analysis, so again not usable for standard engine tests.

And then out of the usable positions in big databases of composed studies, not so few of them are already in one of my databases, like in the one, I started the search for here and now, you see? Same with the 20 to 30, Jouni wrote about those, he found in your github- .epd, they are to a good part already long and well known anti engine puzzles also, searching this forum with expression fortress would bring along many of those as well probably. Some of those then again were too difficult for an automatic screening by engine only finally yet too, if Stockfish doesn't find best move of such really tricky fortress- positions in 30 or 60" single threaded or with many cores SMP at will, doesn't mean it wasn't true fortresses, so not even the final process of filtering is to be fully automated, is it?

And without the prefilter of already collected puzzles, out of which most drawn ones will be fortresses (what should be the "trick" of by single best move forced remis otherwise?), if you meant to simply use positions of low material on board, e.g. out of (human or eng-eng) games, if you filter such positions simply by engine- eval 0.00, how many of these positions will simply be trivially drawn, not by sinlge best move only and not of any difficulty to be evaluated drawn correctly? And how many of even somewhat tricky fortresses will you find that way?
Or did I get you wrong about this one automatic process, you wrote about?