I'm assuming he's marking to market every day, and calculating sharpe on daily basis incorporating daily dd. If he isn't.... Then what's the point.
I understand as a CTA, performance metrics reported by websites like barclayshedge, autumngold etc. are all calculated using monthly data. So sharpe for a CTA would be calculated using month end data, not daily data. Heech your comments please. thanks
Unless your strategy is quirky and extremely non-normal, then really it doesn't matter what time frame you're calculating Sharpe at. It should all roughly be similar, as long as you normalize to annual terms (using *sqrt(T)). (For example, I calculate with both weekly and monthly returns... and the numbers are almost the same.) On the various websites, you're reporting performances monthly... so it makes sense to calculate Sharpe monthly as well. But since the OP has been updating his Sharpe here on much more frequent intervals (random intervals?), then I assume he's calculating it using daily data and then converting to annual terms.
I use daily calc, calculating daily is especially important when measuring DD peak to valley. (my monthly Sharpe is around 7.5 YTD)
yes under normality, equivalence of sharpe calculated using pnl sampled at different times makes sense, just wanted to confirm monthly reportings by CTAs. Thanks!
Yes, it is monthly because that is the required/accepted format for CTA reporting in the DDOC. To be clear, there is no required reporting for risk metrics like Sharpe, Calmar, etc... I am referring to the performance capsule.