Applied Ultrafinitism OR Amending a Forgotten Parameter

December 14, 2011 11:33 PM

S. Ekhad , E. Georgiadis, D. Zeilberger.  "How to Gamble If You're In a Hurry."  arXiv:1112.1645v1 (pdf), 7 Dec 2011.

"The beautiful theory of statistical gambling, started by Dubins and Savage and continued by Kelly and Breiman has mostly been studied under the unrealistic assumption that we live in a continuous world, that money is indefinitely divisible, and that our life is indefinitely long. Here we study these fascinating problems from a purely discrete, finitistic, and computational viewpoint, using both symbol-crunching and number-crunching."

In short, Georgiadis and Zeilberger amend the Kelly criterion with a bound on the time spent in the casino, providing a dynamical programming algorithm for computing optimal bets, but designing this algorithm gets tricky.

If you aren't up for the mathematics, you might still find the differing conclusion on page six of the article's pdf amusing, ditto Zeilberger's opinions.

A few paradoxes to which Zeilberger alludes in the BBC Horizon excerpt include the Borel–Kolmogorov paradox, the Banach–Tarski paradox, the Hausdorff paradox, or really any modern paradoxes. All these paradoxes have infinitary resolutions considered satisfactory by virtually all mathematicians, but that doesn't preclude stranger resolutions, such as simply rejecting them through finitism.

posted by jeffburdges (3 comments total)

Finally prompted me to download Maple for the iPad. Very neat stuff.
posted by Lame_username at 5:02 AM on December 15, 2011

It's worth mentioning that the first author is the name of the third author's computer.
posted by wittgenstein at 8:26 AM on December 15, 2011 [1 favorite]

Specifically, Zeilberger had a 3B1 at some point; he is Israeli; "Shalosh" and "Ekhad" are Hebrew for "three" and "one" respectively.
posted by madcaptenor at 8:38 AM on December 15, 2011