I'm backtesting a strategy that involves monthly investments in a few stocks out of a given set, that is, each month some of the stocks are shortlisted from an index and a long position is taken in them. Each month, the portfolio is equally-weighted in the stocks. The number of trials required for this backtest is $36$ and the backtest period is $84$ months.

Now suppose that instead of an equally-weighted portfolio, each month I construct a Markowitz-optimized portfolio instead. So given any month, optimal weightages to be assigned to the shortlisted stocks are determined and the investment is made in a portfolio constructed on the basis of these optimal weightages. Also the number of steps in each Markowitz optimization is $1000$.

My question: should the number of trials be updated to $36 \times 84 \times 1000$? I'm not clear on this because I tried Googling "markowitz optimization number of trials" and didn't come across any page that discusses treatment of a Markowitz optimization procedure step as a "trial". Would this be too many trials, and how do I know what's a good upper bound on the number of trials?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.