Software for solving chess moremovers

[msterkowiec]

I would like to invite you to familiarize yourself with a program dedicated to solving chess moremovers. You can download a free version at jgisland.pl. It has an extensive test suite that can be viewed at https://jgisland.pl/download/reports/mo ... uthor.html (it is a small database of chess composition). The most recent developments (compositions addressed) are enumerated in readme file: https://jgisland.pl/download/readme_EN.txt . The main goals of the program are speed, robustness and simplicity. I do hope you will like it. Your feedback will be highly appreciated.

[msterkowiec]

Under this link https://jgisland.pl/download/JGIslandBrochure.pdf you can find information brochure of the program - it describes main features and contains a lot of useful information about this piece of software

[Joerg Oster]

Under this link https://jgisland.pl/download/JGIslandBrochure.pdf you can find information brochure of the program - it describes main features and contains a lot of useful information about this piece of software

After a first quick glance I must say this looks impressive.
Seems I can quit my work on Matefish https://github.com/joergoster/Stockfish/tree/matefish2

Seriously, nice to have another mate-solver.

[msterkowiec]

@Joerg Oster
I hope we will be able to inspire each other and share ideas. For sure immediately you may try to reuse the extensive test suite (it took me years to create it; it is still growing).
Here is a full version of the test suite that includes not only moremovers but also two- and threemovers:
https://jgisland.pl/download/reports/re ... yYear.html (it is a quite large html, so it takes a few seconds to load fully)
It contains a column "expected solution(s)", so it should be quite easy to reuse.

[Jouni]

I look same positions, that take long time to solve. ALL positions were solved instantly with Huntsman .

[msterkowiec]

@Jouni
You must have devised some disruptive technology... Could you share any convincing information (links, binaries, reports, etc.) ?

[Dann Corbit]

Bit of an unusual chess set you've got with the following:

[d]7k/8/6Q1/5ppp/4pppp/4pqnn/2p1pqnr/2Kbrrbn w - -
[d]7k/8/6Q1/8/5ppp/4pppp/2p1pqnn/2Kbrrbn w - -
[d]8/6pp/8/4K2n/p3N1p1/rp3prp/1p3ppn/bR3bqk w - -
[d]8/p2p1p1p/p4p1p/5p1p/5p1p/p4B1p/kp1K3p/b6n w - -
[d]B1B4N/1B1B2k1/B1B5/1B2K3/B5BP/5B1P/4B1BP/3B1B1B w - -
[d]B1B4N/1B1B2kP/B1B5/1B6/B5B1/5B1P/4B1BP/K2B1B1B w - -
[d]Nb1QN3/1pp1PPpp/1rnn1p1P/1Pkr2B1/P1pRpK1p/2P1P2B/1Pqb2Pp/4R3 w - -
[d]rbkb1b1b/b1b1bPb1/KbRb1b1b/b1b1bPbB/rb1bRbPb/brb1brb1/1brb1b1b/brqrbrbr w - -

[msterkowiec]

Yes, the intention was to test the program in the usual, unusual and very unusual positions - to make sure it behaves properly in all such settings. BTW: the third position you mentioned:

[d]8/6pp/8/4K2n/p3N1p1/rp3prp/1p3ppn/bR3bqk w - - 0 1

is not that unusual. Its general type is usually called "sea snake". It is dual zugzwang position type - both white and black pieces are in zugzwang. White get out of zugzwang by king manoeuvres (looking for a small triangle to freely turn back to lose tempo). On the other hand, when black pendulum piece (in this case bl.Rh3) cannot move (because of wh.Kb4), Black have to move their pawns (and they ultimately got blocked). This position type has many beautiful compositions, I will mention a few of them:

1)
[d] K7/p3p1p1/1p2p1p1/rp4B1/brpn4/Rnk5/8/4N1N1 w - - 0 1
#163, Ottó Titusz Bláthy, 1890, link1, link2

2)
[d] n1K5/b1p3pp/1pP3p1/1P6/6p1/1p4Pn/1P2N1br/3NRn1k w - - 0 1
#182, Karl Fabel, 1952, link1, link2 (here black bishop is a pendulum)

3)
[d] rBb5/P2p4/P2Ppp1p/1B1p1p1B/2prn3/KRnk4/3p4/b4N1N w - - 0 1
#210, Ottó Titusz Bláthy, 1890, linl1, link2

Famous Hungarian engineer Ottó Bláthy (1860–1939; e.g. co-inventor of the modern electric transformer) was one of the pioneers of chess moremovers and a major contributor as far as sea snake composition type is concerned.

Note also that J.G.Island lets you search compositions by keywords/tags (Ctrl+F). In the standard installation of J.G.Island there's a file sample_compositions.cp, that you can browse for example for a keyword "sea snake" - there are more than 70 such compositions there.

[Joerg Oster]

@Joerg Oster
I hope we will be able to inspire each other and share ideas. For sure immediately you may try to reuse the extensive test suite (it took me years to create it; it is still growing).
Here is a full version of the test suite that includes not only moremovers but also two- and threemovers:
https://jgisland.pl/download/reports/re ... yYear.html (it is a quite large html, so it takes a few seconds to load fully)
It contains a column "expected solution(s)", so it should be quite easy to reuse.

Well, the FENs in this table are incomplete. You should at least provide the info which side is to move.
Or is it always white to move?
Castling flags, 50-move rule and full-move numbers probably don't matter.
Most convenient would be an epd file, for others probably as well.

In general, shorter mates like 2 -6 movers, are no problem for Matefish's Alpha-Beta search.
Longer mates are sometimes solved rather quickly by the optional Proof-Number search.

However, Matefish ist still missing certain knowledge, for example mating with the queen against the lone king.
Note: Matefish is not using any evaluation, Everything is purely done by the search (Alpha-Beta or Proof-Number Search)
and by trying to sort the moves in a most helpful way.

[msterkowiec]

@Joerg Oster
J.G.Island is for orthodox moremovers, so it assumes e.g.:
* White on move
* any castling that seems possible is possible (the assumption is that there were no earlier moves by king(s) and rook(s))
* no retro analysis (no en passant in the first move is considered)
* right, 50-move rule is not considered either

Thus, only position (partial FEN) and stipulation (#n condition) are needed to fully describe a moremover position.
(BTW: Writing a small script, e.g. in Python, for conversion of such html into any format you'd want to use shouldn't be a problem)