A neat twist to the standard pro-rata division we see a lot in bankruptcies.

What I like best is this method penalizes the creditor for extending too much credit. Note that the ones who are hurt most by the divide-contested-amount method are the creditors who over-estimated the debtors assets, while the creditors lending smaller amounts get a small bonus over pro-rata.

Edit: I think the (estate_size == 150) case sucks for the 200 and 300 creditors, though, because they correctly believed the estate would be worth more than 100. It is "pairwise-consistent" as the article defines it.

Here's the discussion from the last time this was submitted:

http://news.ycombinator.com/item?id=461076

Poker players would instantly recognize the similarity of the money splitting problem to negotiating a deal at the final table of a poker tournament (say final 2, 3, 4). It isn't always about dividing the payouts proportional to size of your stack and there are human factors involved.

Great article, but I couldn't help thinking the whole time that while game theory certainly and apparently can be used to describe the split of the estate, so could simple pattern-finding: the contested sum(s) is(/are) split evenly, and the uncontested amount (if any) is given to the one owed the most money.

But perhaps game theory makes the pattern more evident, Regardless, it was a good read.

This is a pretty lame 'religious mystery' in my view. There doesn't seem to be any real wisdom in the method of splitting the debt and the explanation seems fairly straight forward. It isn't even new. Why people find this interesting is beyond me. What's next 'A numerical analyses of the Book of Revelations and Daniel'?