Historical Volatility and Massive Stock Swings


Lets say you’ve been asked to calculate the Sharpe ratio using price data for the last 3 years. In your time-series, the company you’re interested in had a massive spike of 1,250% in one day which lasted a couple of days before mean-reverting (Ticker: SPI). Obviously, if you include this spike in your calculation of volatility, the figure would be grossly inflated and would understate your Sharpe ratio. How would you go about normalizing historical volatility. Would you be comfortable using a vol number that includes this massive spike in your Sharpe ratio?