Community Contributions

[quantopian]

Community Contributions

[BrandonStaab]

Improving the annualization of the Sharpe or Sortino ratios would be of great significance

https://alo.mit.edu/wp-content/uploads/2017/06/The-Statistics-of-Sharpe-Ratios.pdf

[jonathanng]

@BrandonStaab, what is wrong with the annualization as it is?

[BrandonStaab]

@jonathanng

From the paper:

"... monthly Sharpe ratios cannot be annualized by multiplying by sqrt(12) except under very special circumstances, and I derive the correct method of conversion in the general case of stationary returns."

The assumptions being made don't hold up in practice and leads to both over estimation and under estimation of the predicted Sharpe ratio. Each trading period is almost certainty not iid and thus assuming they are might be good enough, but can lead to poor approximations of the true ratio. Correcting for autocorrelations is still just an approximation as the true Sharpe ratio is an unobservable statistic.