I have been running a stock portfolio which iam balancing using markowitz in a custom written mixed integer solver. It lowers the risk and fluctuations a great deal whilst still having good return. What about using this to create a balanced for example naked put portfolio? Anyone tried this? Br Gustaf
Using Markowitz optimization on a portfolio of options sounds too scary. There are more dimensions of risk to worry about besides just movements in the underlying. Also, while stock returns themselves don't actually follow the normal distribution, option portfolio returns REALLY don't follow the normal distribution, and this may violate an important assumption in your model. Maybe the main objection goes something like this: Intuitively, Markowitz type models rely on portfolio variance for quantifying "risk", so that risk and return can be compared somehow. For a portfolio of options, the variance of returns can be a terribly misleading measure of your risk. So at the very least, if you are going to go down this path, find a new measure of risk to put in the model. kps
I don't think you'll be able to apply what you're doing to a portfolio of convex products very easily.
Thanks for the inputs both. I will try it out abit with my paper trading account (IB). Btw here is the correlation of 5 stocks using 30 day history; Corr [*,*] : COKE GE MSFT NLY ORCL COKE 1 0.507973 0.384619 0.413526 0.601868 GE 0.507973 1 0.423308 0.278261 0.675591 MSFT 0.384619 0.423308 1 0.390332 0.836941 NLY 0.413526 0.278261 0.390332 1 0.525658 ORCL 0.601868 0.675591 0.836941 0.525658 1 ---------------------------------------------------- My Idea was to minimize the correlation and ROI % or the probability of gain.
My account awaits funding so I cannot pull historical contract data, but I will also try out running the optimization on the actual historical price of the option contracts and see where that leads me. Any thoughts on this approach? Br Gustaf
I have a formula that calculates the probability of win using volatility, strike price and days left to expiration.
I just realized on thing if markowitz (which works well for my stock portfolio) then also a selling covered calls against this portfolio would work well. Since a covered call is the equvivalent payoff diagram as a naked put, then this would work on optimizing a naked put portfolio as well.. Something wrong in my thinking?