The Kelly Criterion

In an upcoming article at RealMoney, I will be writing an article on optimal position sizing. To do that, I use the Kelly Criterion, which says that a position size should be equal to (edge/odds). There is added complexity in applying this to stocks, because in gambling, each game is uncorrelated with the last one. In investing, if you have a number of investments going at the same time, they are to some degree correlated with one another, particularly for me, because I concentrate sectors.

The Kelly criterion applied to stock investing would recommend a fixed weight portfolio. Optimally, you would rebalance daily to those fixed weights, but there are two factors that interfere: first, there are costs to trading, and second, momentum tends to persist in the short-run. To me that implies that one has fixed weights, but that you set a band around those fixed weights for rebalancing. I use a 20% band, but the more I think about it, the band should be smaller, like maybe 10%. My portfolio has gotten bigger over the past few years, and trading costs are a smaller percentage cost factor. I’ll stick with 20% for now. It has served me well, but I will re-evaluate this.

I firmly believe that my eight rules tilt the odds in my favor. How much are each of the rules worth? Well, that I will describe in the article at RealMoney, and that hopefully will explain why holding 35 or so positions is proper for me in running my strategy.

One more thing, I appreciate the work done at the CASTrader Blog on the Kelly criterion. We manage money very differently, but we both appreciate the value of the Kelly Criterion.

PS — As an aside, I get ribbed by some at RealMoney, whether contributors or readers about holding 35 positions as hyperdiversified. In general I agree with portfolio concentration, but given that I concentrate sectors and industies, that makes 35 a lot more like 20 in terms of total volatility. On the other hand, compared to most mutual fund managers, 35 names would be a tighter ship than 95% of them run.