4 ms·
Kelly criterion is the solution for optimal bet size with the assumption that utility is a log function. While it's often modelled as such we know it's not per
by bluecalm 5y ago
Kelly criterion is the solution for optimal bet size with the assumption that utility is a log function.
While it's often modelled as such we know it's not perfect (utility of money is bounded or even goes downwards after certain point while log is unbounded) and more importantly utility of money is different for different people.
That means that, while being an important illustration of the concept of utility of money, Kelly is most likely not a good practical guideline for most people. Many professional gamblers for example feel that something closer to 1/2 Kelly or 1/4 Kelly is better but again it all depends on your personal situation and what money means to you.
- kqr 5y agoAnother comment already posted this misconception. The Kelly criterion assumes only two things: compounding and that the objective is maximised growth. It makes no assumptions about utility functions. There are several good reasons to use fractional Kelly strategies, but shapes of utility functions has nothing to do with it.
- bluecalm 5y agoHow do you define maximized growth and why is it important for anything if utility is not log? Shapes of utility function have everything to do with determining your bet sizes. Can you provide one reason to use Kelly that is not connected to the shape of the utility function? Not to escalate but calling a very well known fact about Kelly a misconception doesn't make for a good start. I mean it's math you can verify or open Wikipedia and read the very first paragraph there.