I'm surprised only three solved all the test data - could you at least publish some selected test data which caught out the majority? Unless you were using some weird input format cases, I assumed most would spot the "designed-to-catch-you-out" cases (ie question 1 - anything which overflowed an int, question 2 - anything which had a non-"greedy" optimal).
I assumed we'd get to see the test data after - otherwise there's not much incentive to enter. I imagine most didn't expect to win, but were entering for the challenge of spotting your sneaky boundary cases - it's a bit disappointing to be told "most of you failed to pass the test" and then not hear what we failed at! I guess I won't be rushing back for the next one.