Calculating Sharpe ratio for intraday futures strategy?

Discussion in 'Automated Trading' started by JangoFolly, Aug 8, 2005.

  1. I understand the calculation of the sharpe ratio as it relates to long-term investing of non-margin positions (its most basic implementation). I'm developing an intraday trading system for YM, ER2, and ZN, and I'm a bit lost in what values to stick in the equation. Should the percentage return be based on the intraday margin amount for each contract or some other value? I'm guessing the risk-free return could be represented by the sweep return on my account; however, I earn that return when I'm trading as well as when I'm not (i.e., it's a constant rather than an alternative to the return on the trade). Would it be wrong to calculate the ratio as the mean of the return divided by the standard deviation of the return?

    Also, I'm currently trying to balance profit/loss ratio versus trade frequency. What profit/loss ratio levels do you consider to be acceptable, good, and excellent?

    Thank you.


    Regards,
     
  2. Sending back up to the top of the list...
     
  3. One last try...
     
  4. Hello:

    Interesting that no one responded to this simple question. Seems to say a lot about the education and skill level of the crowd here at ET. I'm not going to say that they are a bunch of idiots, but the thought does cross my mind.

    First, you need to understand that Sharpe ratio measures the return IN EXCESS of a guaranteed investment, relative to its risk. In theory you can use any guaranteed investment, but actually you should be using something like the return on the 90 day T-Bill.

    If you want to calculate a Sharpe ratio of your trading program. Simply take the return for some period of trading, and divide that portion of the return in excess of the 90 day T bill, by the standard deviation of those returns. Now that I think about it, you may want to annualize your returns to make it easier to calc.

    As an example, if your trading produced an annualized return of 25%, with a standard deviation of 10%, while at the same time 90 day T-Bills returned 5%, your trading program would have a Sharpe Ratio of 2.


    25% minus 5% divided by 10% = 2.0


    The higher the ratio, the better the return per unit of risk


    Edit:

    On second thought, I just read some bullshit comments from people who can't seem to get the message that education and hard work are important to trading success. Now that I think about it, they are (with some exceptions) frigging idiots. :D

    Steve
     
  5. Chagi

    Chagi

    Two things.

    First, to supplement the previous post, here is the formula for Sharpe's Measure, straight out of a textbook:

    (Rp - Rf)/Std. Dev of P

    Rp = average portfolio return
    Rf = average risk free freturn

    The above two are divided by the standard deviation of the returns for the period.

    Second, judging by the books that I've read so far, your win/loss ratio isn't nearly as important as your money management, i.e. letting winners run and cutting your losses short ASAP. You could probably succeed if you were right only half of the time (possibly) less, so long as each loss was small and each win was larger than your losses.
     
  6. I think it should be based on the account size required to trade the system, not on intraday margin requirements. That is the amount of money tied up in trading the system.
     
  7. goliant

    goliant

    It isn't good to use Sharpe's ratio for intraday strategies. Because individual intraday trades tend to have small returns, the standard deviation of all these trades will be small as well. Therefore when dividing by the standard deviation to calculate the ratio, the result is larger sharpe's ratios compared to strategies with the same overall return but with fewer/longer trades.
     
  8. mhb

    mhb

    If I was you, I would take your total return for each day.
    After about 50 days (statistically significant) you can calculate a standad deviation of those daily returns and base your sharp ratio on that figure.

    Daily returns are a minimal time frame, as interest rates generally aren't calculated by the hour.
     
  9. cakulev

    cakulev

    Well it all depends how you calculate return. Normally you use return on equity so this is what you use. The risk free return is the one based on the equity not capital. Otherwise you would have 0 or negative Sharpe, even though you had 100%+ return.
     
  10. Calculating a margined instrument, say future is quite standard in Sharpe ratio calculations:

    http://www.miapavia.it/homes/ik2hlb/sr.htm

    For an intraday, I would use your intrday Maximum position as contract sized to be used in the calculation (for margining, etc), then calculate the daily returns, and so forth.

    I agree that 90 day T-Bill should be used as the risk free rate. I calculate my sharpe ratio every month, using daily return data.

     
    #10     Aug 20, 2005