Due to the new filters slowing things down, and since there are many design puzzles up, we've extended this deadline by 1 week.
FYI: I ran mutate for 8 hours with filters on, and at the end, I have "Too many polar residues in the core" and 19 invalid residues. (monomer is all helices).
Just want to report to the devs that once again, even with the filters, a megahelix scores far higher than any realistic structure.
You can find it here: http://fold.it/portal/node/994899
A summary about this issue is that in symmetry these helices are boring, but they're stable. So it's not super surprising that they score well.