1) Yes, Michael Keller wrote this to me when I first joined the group.
I agreed and said that I should have worded my request better.
2) Michael Keller made this suggestion after reading my reply to (1).
He suggested that I might find long solution problems by examining deals
that couldn't be solved with three freecells. He also told me how to get
FcPro26 to run with only three freecells and get it to list only those
that couldn't be solved. I did this for the first one million deals. I
then wrote a converter for PatSolve so that I could count the distinct
moves in its output. I then took the results from FcPro26 and ran each
of them through PatSolve to solve with four freecells. As each run
completed, the results were passed to my converter and I saved the
number of direct moves it took to solve each. Some of the deals took
only a modest number of moves to be solved with four freecells. Others
took several more moves. For each deal that took PatSolve 69 direct
moves or move to solve with four freecells, I ran that deal through my
solver to see if it could solve them and to see how many (near-)optimal
moves each would take. Ironically, there was only a moderate correlation
in the number of moves between PatSolve and my results.
3) Once I had learned that Freecell Solver could be put into "Fool's
Gold" mode to produce atomic moves, I tried to write a converter for its
output as well. I encountered problems and documented them in an earlier
message to the group. After not hearing anything in response, I
discarded my converter effort.
Regards,
Danny A. Jones
Shlomi Fish wrote:
>
> This was discussed here a while back. One of the conclusions that were
> reached were that:
>
> 1. The length of the shortest route usually does not indicate how hard a
> game it is, at least not according to a human player. Mike Keller noted
> that some of the most difficult deals in his collection of MS Freecell
> solutions have very short solutions, while some very simply ones
> have very long ones.
>
> 2. One way to approximate the difficulty of a deal is to see how many
> Freecells are required for solving it. The more Freecells are needed, the
> harder the deal is.
>
> One thing I can think of right now, is to see how well the deal can be
> solved with only a subset of Freecell Solver's meta moves. I remember that
> an early version of Freecell Solver in which I implemented only some of
> the moves was able to solve some deals I prepared.
>
> Regards,
>
> Shlomi Fish
>
>
> ----------------------------------------------------------------------
> Shlomi Fish shlomif_at_vipe.technion.ac.il
> Home Page: http://t2.technion.ac.il/~shlomif/
> <http://t2.technion.ac.il/%7Eshlomif/>
>
> An apple a day will keep a doctor away. Two apples a day will keep two
> doctors away.
>
> Falk Fish
Received on Sat Nov 29 2003 - 18:59:33 IST