Asymmetrical Information

From the desk of Mindles H. Dreck:

Counterintuitive

In discussion of the post below, Jonathan Gewirtz observes:

But you can reduce "path risk" by reducing bet size.....a participant in the experiment is likely to lose money if, as Paulos assumes, he bets his entire stake on each trade, because in that case it takes only a few losses in a brief sequence of trades effectively to wipe him out. But in real life you don’t have to bet that big. Your odds of long-term profitability are much better if you risk only a small fraction of your total capital on each trade. Bet-size optimization can't transform a losing game into a winning game, but you can blow up in a winning game if you bet too big. Small systematic variations in bet size can create enormous differences in end-state wealth in positive-expectation games.

True, but it doesn't improve the strategy, as you point out. However, the comment about "transforming a losing game" reminds me of Parrando's Paradox, which shows that you can actually combine two losing games into a winning game. How odd is that? John Allen Paulos has covered this one as well.

Posted by Mindles H. Dreck at April 25, 2002 6:14 AM | Technorati inbound links