My problem may be in wanting to "compare apples and oranges". However,
counting the WKR automoves probably gives it a disadvantage against
Horne automoves when comparing short solutions for the same deal.
I finished investigating 32 deals with short solutions derived using WKR
automoves and not counting the automoves. When most of these deals were
"exploded" to show the WKR automoves, I noticed that some of the WKR
automoves were extraneous -- especially among those near the end of a
solution. This translates into an "exploded" WKR solution being longer
than its Horne counterpart for these deals.
However, my solver does find solutions quicker using WKR automoves. So,
I'm keeping it active in my solver for now.
Separate (sic?) note: I just discovered that FcPro doesn't allow a
multi-card move when the smallest card in the chain is an Ace. Something
I don't recall encountering previously.
Regards, Danny
--- In fc-solve-discuss_at_yahoogroups.com, "Gary Campbell" <gary_at_...>
wrote:
>
> Probably 27 until WKR becomes the standard for Automoves. However,
when I benchmark my solver I use WKR to tabulate the minimum, medium,
and maximum solution lengths of games 0-999. Always making the WKR move
seems a reasonable thing for a solver to do, whether it’s counted
or not.
>
> If you want solutions that play out on popular FreeCell Players, you
probably have to stick with Horne Automoves.
>
> One final thought: There is probably a difference between Horne and
WKR with respect to shortest solutions. I can imagine two games such
that using Horne vs. WKR, game A has a shorter solution than B vs. a
longer solution than B, if you catch my drift. And this could occur
either way. It’s harder to imagine a WKR solution being longer
than a Horne solution for any given game, however.
>
> -Gary
>
>
Received on Mon Oct 22 2012 - 14:51:04 IST