[Retros] A new program to solve fairy proof games
caillaud1957 at gmail.com
Sun Nov 5 03:47:23 EST 2017
Oups, so my diagnosis was not "obvious" but totally wrong!
I didn't suspect the innocent-looking and "natural" 0-0 as in Volage and
Hypervolage, en passant and 0-0-0 are tricky and "not so natural" moves, so
I imagined that a computer-program can go wrong there.
-in en passant, the "transitory" change of color is ignored (a useful trick
to avoid duals in helpplay...)
-0-0-0 is considered as a King move so the Rook doesn't change color
On Sun, Nov 5, 2017 at 6:36 AM, François Labelle <flab at wismuth.com> wrote:
> Hi Michel,
> Thank you for the bug report.
> Actually the bug has to do with castling and not en passant. A minimal
> example is
> forsyth rsbqk2r/ppP1bppp/8/8/8/8/PPP1PPPP/RSBQKBSR
> stip a=>b 1.0 forsyth rSbq1rk1/pp2bppp/8/8/8/8/PPP1PPPP/RSBQKBSR
> cond Hypervolage
> Jacobi v0.1 fails to see the solution 1.cxb8=S O-O.
> This bug affects proof games with a condition where a piece can change
> color, so Andernach, AntiAndernach, Volage, Hypervolage, Masand, and Circe
> Turncoats. Such problems that were verified using Jacobi v0.1 shouldn't be
> considered C+ because cooks with castling can be missed. The bug doesn't
> affect Tacu enigmas (proof games with only undefined pieces and
> ColorThePieces). The bug will be fixed in v0.2 which I think will be ready
> sometime this week. Sorry about the bug!
> CAILLAUD Michel wrote:
>> Going on testing version 0.1...
>> Jacobi 0.1 doesn't solve PDB P1240567
>> Guy Sobrecases dia5.5 r1bq1rk1/pp1sbppp/8/8/8/8/PPP1PPPP/RSBQKBSR
>> The reason obviously is that Jacobi 0.1 ignores en passant capture in
>> Hypervolage (Popeye and Winchloe play it).
>> So, with this version, Jacobi cannot be used to test Hypervolage problems
>> (cooks with en passant can be missed).
> Retros mailing list
> Retros at janko.at
-------------- next part --------------
An HTML attachment was scrubbed...
More information about the Retros