Advanced search
Quantiacs
Quantiacs > Sharpe ratio calculation View modes: 
cgorac - 1/17/2019 2:39:58 AM
   
Sharpe ratio calculation
Hello,

I was looking into Sharpe ratio calculation segment of code in stats.m (am using Matlab version of toolkit).  I understand the formula, but in most references that I was able to find, annualized Sharpe ratio seems to be calculated simply as follows (when returns vector contains daily excess returns):

sharpeRatio =  sqrt(252) * mean(returns) / std(returns);

However, in looking further for explanation of using sqrt(252) factor for annualizing standard deviation of returns, I found reference to Andy Lo paper "The statistics of Sharpe ratios" (https://www.researchgate.net/publication/228139699_The_Statistics_of_Sharpe_Ratios), where this factor is derived starting from an assumption that returns are log-normally distributed.  In that light, calculation implemented in toolkit indeed makes more sense to me.  Still, anyone else care to comment on this discrepancy?

Thanks.

1