Software for solving chess moremovers

[msterkowiec]

Version 9.2 of J.G.Island is now ready for download at https://jgisland.pl.
There is a set of relatively small improvements, except for the first one:
a) improvements for positions wh.Q vs. bl.pawns that were necessary for successful solving of #118 8/5p1p/2p4p/2p2n2/p3Q2p/K6p/2r5/1k5n https://pdb.dieschwalbe.de/search.jsp?e ... 'P1183413' (now ~1h)
b) improvements for brute force (changes in subroutine SortBlackMoves)
c) initial improvements for positions that are not a typical moremover composition e.g.: 2b1k3/1p2b3/p3P1pp/2pP1p2/5B2/2NQ1N2/PPP2PPP/2KR1B1R (#6) or 2b1k3/1p2b3/p1p1Pppp/8/3PQB2/2N2N2/PPP2PPP/2KR1B1R (#9)
d) (GUI) menu option to search in Meson database

The improvement for position type wh.Q vs. bl.pawns was a special one: it required a record number of more than 120 unit tests on SortWhiteMoves subroutine. It turned out to be the only way to make the program work properly in this position type (no tricks with draw evaluation, no reliable heuristics, etc.)

[F.Huber]

2b1k3/1p2b3/p1p1Pppp/8/3PQB2/2N2N2/PPP2PPP/2KR1B1R (#9)

Hmm? That's #3, not #9.

[msterkowiec]

My bad, this position was actually meant:
2b1k3/1p2b3/p1p1Pppp/8/3P1B2/2N2N2/PPP1QPPP/2KR1B1R

[F.Huber]

My bad, this position was actually meant:
2b1k3/1p2b3/p1p1Pppp/8/3P1B2/2N2N2/PPP1QPPP/2KR1B1R

Ok, but also this one is #7 instead of #9:

ChestUCI Ver.5.2:
FEN: 2b1k3/1p2b3/p1p1Pppp/8/3P1B2/2N2N2/PPP1QPPP/2KR1B1R w - - (16+9)
Stellungs-Analyse: C0/R0/K3/P9/X17 W:12/44
AutoTurbo-Suche nach Spezial-Matt [C1/R0/K0/P0/X0] in 9 ... (Hash=256MB)
AutoTurbo-Suche nach Spezial-Matt [C0/R1/K0/P0/X0] in 9 ... (Hash=256MB)
AutoTurbo-Suche nach Spezial-Matt [C0/R2/K0/P0/X0] in 9 ... (Hash=256MB)
AutoTurbo-Suche nach Spezial-Matt [C0/R-2/K0/P0/X0] in 9 ... (Hash=256MB)
AutoTurbo-Suche nach Spezial-Matt [C0/R3/K0/P0/X0] in 9 ... (Hash=256MB)
AutoTurbo-Suche nach Spezial-Matt [C0/R-3/K0/P0/X0] in 9 ... (Hash=256MB)
AutoTurbo-Suche nach Spezial-Matt [C0/R4/K0/P0/X0] in 9 ... (Hash=256MB)
AutoTurbo-Suche nach Spezial-Matt [C0/R-4/K0/P0/X0] in 9 ... (Hash=256MB)
7/9 00:07 478.322 567.404 +M7 1.Qe4
7/7 00:07 4.665.756 807.084 +M7 1.Qe4 f5 2.Qe5 Bxe6 3.Re1 g5 4.Qxe6 Kf8 5.Qxe7+ Kg8 6.Bc4+ Kh8 7.Be5+
AutoTurbo-Suche nach Spezial-Matt [C0/R-4/K5/P0/X0] in 7 ... (Hash=256MB)
7/7 00:12 466.520 574.532 +M7 1.Qe4
7/7 00:12 4.438.343 821.003 +M7 1.Qe4 f5 2.Qe5 Bxe6 3.Re1 g5 4.Qxe6 Kf8 5.Qxe7+ Kg8 6.Bc4+ Kh8 7.Be5+
AutoTurbo-Suche nach Spezial-Matt [C0/R-4/K5/P0/X27] in 7 ... (Hash=256MB)
7/7 00:17 461.446 568.283 +M7 1.Qe4
7/7 00:17 3.833.349 823.313 +M7 1.Qe4 f5 2.Qe5 Bxe6 3.Re1 g5 4.Qxe6 Kf8 5.Qxe7+ Kg8 6.Bc4+ Kh8 7.Be5+
Matt in 7 gefunden ! (1 Lösung in 00:04)
7/7 00:17 3.833.349 823.313 +M7 1.Qe4 f5 2.Qe5 Bxe6 3.Re1 g5 4.Qxe6 Kf8 5.Qxe7+ Kg8 6.Bc4+ Kh8 7.Be5+
Best move: Qe4, Value: +Mate in 7, Depth: 7/7, Time: 00:17,198, 3.833.349 Nodes, 823.313 N/sec

[msterkowiec]

Well, right, anyway in this case this is actually irrelevant : ) In my case this position was a subproblem of some other problem and #9 was needed in this subproblem. So this small development was to make this position to be solved efficiently and tests were made assuming #9...

This way or another, as you can see this part of changes for version 9.2 (part c) is marked as "initial improvements", so this is only some partial effort to make this rather untypical position type to be handled in an efficient way...

BTW: The main position is this one:
https://forum.computerschach.de/cgi-bin ... #pid138997