---In fc-solve-discuss_at_yahoogroups.com, <shlomif_at_...> wrote :
Hi Theodore,
On 11 Jul 2016 03:31:41 -0700
"theodorepringle1_at_... mailto:theodorepringle1_at_... [fc-solve-discuss]"
<fc-solve-discuss_at_yahoogroups.com mailto:fc-solve-discuss_at_yahoogroups.com> wrote:
> Hello everyone,
>
> I've been working on various FreeCell projects for more than a year. A few
> months ago, I started my attempt to solve by computer all FreeCell Pro deals
> with four freecells, and to find all of the impossible deals. I have finally
> completed the computation, and I found 102,032 impossible deals, as well as
> 46 intractable deals. Here is the list:
> https://drive.google.com/folderview?id=0B6WZWNModsLfaEIzbnM5UkpHek0&usp=drive_web https://drive.google.com/folderview?id=0B6WZWNModsLfaEIzbnM5UkpHek0&usp=drive_web
>
Thanks for sharing - I'll take a look.
> I used Tom Holroyd's solver in FreeCell Pro. The newer versions of Shlomi
> Fish's solver are much faster and I would have liked to use it, but I found
> that its deal numbers only match FreeCell Pro's from deals 0 to 2147483647,
> and then it cycles between those deals, so its deal number 2147483648 is the
> same deal as deal number 0. I was able to speed up the computation process by
> solving different ranges of deals on multiple copies of the program, and I
> also had two computers to do the work on.
I think the deal generation of Freecell Solver was amended to handle the
extended deal range, but it may only be in the git master version - I an
check.
>
> I found 48 deals which could not be resolved by Holroyd's solver using 2 GB
> of memory. The first two such deals, 53687601 and 186461199, are known to be
> impossible. One particularly curious deal is 2303926468, which I could not
> definitively resolve with either three or four freecells.
>
OK.
> I was surprised to find some impossible deals not on Shlomi Fish's list of
> impossible deals, while I found no new winnable deals.
This may likely be due to deals-falsely-reported-as-impossible by Tom Holroyd's
solver which incorporates some prunes that added some false negatives.
I checked some of them with your solver (with the toons-for-twenty-somethings preset) and it finds all of them to be impossible. The deals which are resolved the most quickly seem to be 517280516 and 541528586.
Imp4.txt is the list of deals impossible with four freecells, Imp5.txt is the list of deals impossible with five freecells, MoreImp4.txt is the list of deals Tom Holroyd's solver considers to be impossible that are not on your list of impossible deals, and Int4.txt is the list of deals that are not definitively known to be winnable or impossible.
> I also found some
> winnable deals which are considered impossible by the Woods-Ettlinger solver
> in FreeCell Pro, such as 93877578 and 1129622823.
Sounds likely.
>
> Finally, it definitely deserves to be mentioned that there are six FreeCell
> Pro deals impossible with five freecells. They are 14720822, 647519280,
> 3015622123, 5705339228, 7273316239, and 7573544118. All of them can be won
> with six freecells. It should be noted that 3015622123 is impossible with
> four freecells and one empty column, while 7273316239 can be lost with six
> freecells by playing a bad first move. Those two deals are probably the most
> difficult of all FreeCell Pro deals.
ah.
Thanks for your effort.
Regards,
Shlomi Fish
Received on Thu Jul 14 2016 - 07:24:14 IDT