John, I wonder if my calculation on the file is done properly for the period? Also if it is done how come such a low Sharpe Ratio with what would be considered a good return. Thanks for the help, Tula
Your calculation is correct. Upon closer inspection, it appears that your style shift might have occurred in the third year. Only you can know this for sure. If you were doing things differently, then exclude the data. If you were doing things the same, then do not exclude the data. Do not worry too much about your Sharpe Ratio. It says more about your style than your performance. Arb guys or "option sellers" tend to have high Sharpe Ratios (until the inevitable six-sigma event occurs) and trend-followers or "option buyers" tend to have low Sharpe Ratios. There is lots of literature on the web that you can find discussing the Sharpe Ratio and its strengths/weaknesses. The Sharpe Ratio measures both return and variability of return, regardless of direction. One of the weaknesses of the Sharpe Ratio is that it penalizes large positive returns as much as it penalizes large negative returns. For example, the Sharpe Ratio (1.32) of a trader who returned 30%, 100%, and 30% would be significantly lower than the Sharpe Ratio (2.02) of a trader who returned 30%, 10%, and 30%. Which performance would you prefer? jj
John, Thanks what you mentioned about the same Sharpe ratio for the 30 100 30 being the same as the 30 10 30 is exactly what was confusing me when i was playing with the numbers. I thought that many Hedge Funds looked at this Ratio as a very important ratio for the Risk of having the returns. What are other important Ratios that the Funds look at? Are you involved in managing money yourself? thank you, Tula
Yeah. In fact, 30-100-30 comes out significantly worse than 30-10-30. (1.32 vs. 2.02) There are a hundred different measurements that people use to varying degrees. The bottom line is that there is no "best" or "ideal" measurement of risk-adjusted performance. So far, the best technique I've seen is to adjust all results to the same annualized standard deviation and go from there. Yes, I am a CTA. jj