Article about using the Kelly formula for position sizing

[alexander]

Very interesting article by Michael Harris about expected gain and the Kelly formula titled "Relation of Expected Gain to Kelly Formula":

http://www.tradingpatterns.com/About_Us/articles/articles.html

I would like to hear your comments about the use of the Kelly formula. To me, even 0.5 x Kelly% is way too much of a risk.

Alex

 

[ZDO]

I use the slope of a 4 period Juric moving average of my own equity curve for a system to determine how close to optimum-f to run. In most systems the allowed range is 43% to 102% of optimum-f, but some are constrained to ranging between ~30 to ~70 % of optimal-f . My (not necessarily mathematically sound) criteria for setting these ranges is on that key factor in optimal-f which is the 'largest loss' figure you give it to perculate with.
Note: This kind of mm can still result in some pretty wild and agonizing rides - but at least the recoveries from dd's are quick.

 

[charliechan]

thanks alex. interesting stuff.

here's a link to more kelly stuff for those interested - less mathematical and more on the underlying thought...

http://www.lmcm.com/pdf/sizematters2.pdf

 

[TWI]

I think kelly is a nutcase.