[MTGandP]

> It's my life's dream to do a real-life process vs. algorithm design comparison like this one.

I've done some simple analysis of card sorting. I've timed myself vs. other people sorting cards, where I use bucket sort and they use whatever they want. I usually beat them by about 20%, but that may just be because I've practiced more. I think bucket sort would have a clear advantage over naïve sorting if the deck had about 100 cards or more.

I've also tested bucket sort against merge sort, and they're about the same speed.

For my bucket sort algorithm, I use two iterations. First I sort the cards into three buckets, where the first contains cards 2 thru 5, the second contains 6 thru 10, and the third contains face cards. Then I do insertion sort on the buckets. On mergesort, I switch to insertion sort for sufficiently short lists.