Fw: solver for kpatience spider is sometimes wrong

[Shlomi Fish]

Forwarding with permission.

Begin forwarded message:

Date: Tue, 28 Mar 2017 15:35:52 +0300
From: Shlomi Fish <shlomif at shlomifish.org>
To: "J.J.van.Beele" <jeroen at guts4roses.org>
Cc: Shlomi Fish <shlomif at gmail.com>
Subject: Re: solver for kpatience spider is sometimes wrong


Hi J.J.,

On Tue, 28 Mar 2017 14:10:05 +0200
"J.J.van.Beele" <jeroen at guts4roses.org> wrote:

> hi shlomi
> 
> i enjoy playing kpatience/spider very much
> 
> what surprises me is that the solver sometimes states that the current game
> is no longer winnable whilst it still is, recently i had the following
> experience:
>   

Unfortunately, you've emailed the wrong guy. I didn't contribute to kpat's
spider solver (nor do I play spider a lot). I am the author of
http://fc-solve.shlomifish.org/ which is a solver for Freecell, Simple Simon,
and some similar games. An old version of it used to be integrated into kpat,
but eventually it was replaced by a fork of patsolve (see
https://github.com/shlomif/patsolve and http://cards.wikia.com/wiki/Patsolve
), before it was later relicensed under the Expat licence and continued being
developed by me (and later integrated into FReecell Solver). I think that
kpat's spider solver was originated by Stephan Kulow ("coolo").

Note that I was once emailed about a different solver for Spider that someone
wrote and he gave me the source, but:

1. Its licence was not explicitly specified there.

2. It used many of his custom header files that were not provided as part of
the source.

3. It made a bad use of multi-threading to implement preemption.

---

With your permission I'd like to forward this reply to the kdegames and
fc-solve-discuss mailing lists.

Regards,

	Shlomi Fish


> during playing spider the solver at some point said that this game could not
> be won anymore my next step is to move a pile of cards in a reversible way
> (can't remember exactly, but you should think of moving a queen lying on a
> king to a free column, this is a reversible move in the sense that i can move
> the queen back onto the king) now the solver says that this game is winnable
> 
> strictly logically speaking this is a contradiction
> because if i evolve to a winnable position than my starting position is
> winnable aswell
> 
> my question to you: are you aware of this fact? should i interpret the status
> "is no longer winnable" in a different way (like: i, the solver, can't win
> this game anymore)
> 
> anyway, i'm a mathematician and wouldn't mind looking into this
> i'm busy though, i just write you because i notice this quite regularly
> 
> regards
> 
> jeroen j van beele, olifantswerf 2, 1018 js  amsterdam
> +31(6)48472527, +31(20)7858909; jeroen at guts4roses.org; guts4roses.org 
> LET OP: mickeysoft outlook verwijdert de door mij geplaatste regeleinden
> herstel deze regeleinden zodat je de originele formatting herkrijgt