I suspect that the contest will really end in a few days at most. Just as soon as some one submits the optimal card arrangements its electively over.
Seriously, I think you under estimate the rigour of those posing the questions But do enter, its a lot of fun!
Brute force search for the optimal solution works for up to 13 cards, but not for 97. I'm looking for mathematical structure to the solutions, but there are no obvious patterns in the solutions for small decks that I see.
Funny thing: The ICFP contest is from the university I'm going to.
Nice to see you on the mailing list, Matt.
My brute forcer takes 1.5 secs on N=13 on my laptop, and I've bruted as far as N=19 (almost finished running now, a week later). Brute as in a lot of culling, but no guessing, and exhaustive. I can't go to further N, obviously. Possibly one step further if I borrow a cluster. But that isn't going to make me top-10 in this contest.
And there are plenty of patterns, but the next N that I brute seems to invalidate or add exceptions to the patterns I've inferred from the lower N....
As I know that some of the top-3 are using GWBasic, there is clearly patterns that we haven't yet seen...
If I understand it correctly, the 'pure' topswops is a win/lose thing; from a randomly shuffled pack of cards, solve it until you reach 1, and you win if the pack is sorted. That is to say, there are plenty of solutions where the pack is not sorted. Knuth and Sudborough+Morales set bounds in this search, rather than finding the solution with the most steps. That's if I've understood the links.
There are lots of people who have submitted the known-optimal N=19 in terms of this contest. I was optimistic saying my N=19 brute was nearly finished; its still crunching away, and I know others that are also still running N=19.
But clearly its brains not braun that will win this contest.