In a message dated 12/10/01 8:57:43 PM Pacific Standard Time,
aettlinger_at_worldnet.att.net writes:
> <<Are you writing of moves internal to the solver, or moves expressed in a
> printed solution???
> The expressed moves in the solution that is output to the user. If you
> have a copy of FcPro, you could easily observe the move pattern.
Your earlier statement has not been brought forward, but you wrote that FcPro
produces mostly single-move solutions, rarely two-card moves, maybe never
more. But surely FcPro will employ longer moves when useful, because, as I
quote from Michael Keller's FAQ page: "Free Cell Pro now works correctly in
all Supermove situations."
> < failure to apply its automove function.>>
> This had to do with an ace being in a freecell. I guess the real
> problem was that FcPro never thought an ace would be in a freecell because
> it would never put one there.
Oops! I slid right over that ACE in the freecell. How does a solver put an
ace into a freecell after a long development and intense usage?
> <<I'm DESPERATELY trying to alert all of you that (1)(a,b) is a DISASTROUS
> limitation that MUST NOT be allowed to HANDCUFF us forever.>>
> No, I really do not understand the reason for any excitement on this
> issue. This does not have anything to do with the solver's ability to find
> a solution, does it? Isn't this entirely a matter of "publishing style",
i.
> e., how the solution is to be presented to the user?
MS freecell allows for only two possibilities: move one card, move the
longest legal sequence. Nothing in between. In other words, some perfectly
legal moves are not usable in certain situations. That's OK for MS freecell
and our printed solutions for everyday users, but the freecell solver
community must not be hogtied by that limitation.
Aside: I suggest that it may be possible to create a legal freecell position
that is unsolvable because of that MS limitation.
> < discussed it with others before anyone else thought about it.>>
> What is the nature of the "rule". Is it a formula which yields the
> number of cards that can be moved from a source to a destination, the input
> parameters being the number of empty freecells and the number of empty
> columns?
No. It's an automove scheme, similar to Horne's but stronger and just as
safe. tomh asked the same question first - I'll explain it there. A formula
that calculates the longest sequence that can be moved is easy to work out.
Surely all solvers use one.
> Do you consider it legitimate, to squeeze out a higher yield, to make
> part of the move a shift of a single card to or from a freecell to or from
a
> column? I. e., if one has three and one vacant spaces, changing it to two
> and two allows more cards to be moved.
Given three empty freecells, one empty column and a second column occupied by
just one card, I would consider it not only legitimate but very desirable to
produce a shorter solution by moving the single card from a column to a
freecell, thereby permitting a much longer sequence to be moved. But I would
count the single-card move as a separate move, not as "part of the move."
> The idea of maximizing the number of cards that can be moved by one
> source-destination command is nothing more than providing a convenience for
> the player.
I wouldn't say "nothing more."
> It does not make-or-break the process of finding a solution.
True.
> Whether it really provides a useful shortcut for a solver to use is another
> question.
My solver doesn't specifically look for that situation. Whether it has ever
actually used that trick when looking for a shortest solution, I can't say.
Moreover, my solver doesn't ever look for ANY specific sutuation.
Bill Raymond
Received on Mon Dec 10 2001 - 22:41:37 IST