1
$\begingroup$

I have an intraday strategy, which will place 0-5 trades for each intraday trading session. (Note that some days it will not place any trades out). The average duration of a trade is around 33 minutes.

From my understanding, the sharpe ratio formula is calculated as mean of excess return of all the trades divided standard deviation of excess return of all the trades. $$E(Excess\; Return) / \sigma({Excess\;Return})$$ Excess return is return of a trade ($trade\;side \times (\frac{executed\;price}{cost} - 1)$) minus the risk free rate.

Now to calculate the annualized Sharpe ratio, it's normally multiply by $\sqrt{252}$. My question is would the calculation of sharpe ratio be different for my intraday strategy ? Would I use a different multiplier other than $\sqrt{252}$ ?

Improve this question
$\endgroup$
4
  • 2
    $\begingroup$ I honestly do not understand why a detailed question like this is being closed ... $\endgroup$
    – user1769197
    Dec 6, 2022 at 13:38
  • $\begingroup$ See quant.stackexchange.com/review/close/42382 . Appears to be a split decision, which on most sites would leave a question open. QF seems to close discussion fast and often. $\endgroup$
    – krkeane
    Dec 6, 2022 at 16:44
  • $\begingroup$ Your question seems to have multiple (related) sub questions which is indeed not very focused. Are you able to split of one of the sub questions into a separate question? Note, that you're more likely to get answers on multiple focused questions: people that might be able to answer a part might not do so because they don't want to or can not answer another part. $\endgroup$
    – Bob Jansen ♦
    Dec 6, 2022 at 19:43
  • $\begingroup$ ok removed the sub questions on transaction cost and margin. Please re-open the question $\endgroup$
    – user1769197
    Dec 7, 2022 at 4:42

1 Answer 1

Reset to default
3
$\begingroup$

The Sharpe Ratio is calculated using returns over predefined intervals of time, typically trading days. I do not think returns over each trade (each which lasts a random time averaging 33 minutes) are a valid starting point for calculation of a Sharpe Ratio.

My suggestion is: simulate you trading strategy over a test period of $N$ days. On days when there are no trades, the calculation is easy: the excess return is zero. On other days do a realistic simulation, including rules on position sizing based on your account balance that day (assume you start the first day with 1,000,000 USD), whether you use leverage or not, whether you allow overlap (opening a 2d trade while the first is still under way), what the transaction costs are, etc. Then from the $N$ daily returns compute the Sharpe Ratio in the normal manner.

Improve this answer
$\endgroup$
4
  • $\begingroup$ Would you be able to give an example of how you would compute the daily return if there are some buy trades and some sell trades on the same day. $\endgroup$
    – user1769197
    Dec 7, 2022 at 9:29
  • $\begingroup$ actually, if there are some buy and sell trades same day, can i just compute daily return as cumprod(1+rt) - 1, where rt is the series of returns from each individual trades ? $\endgroup$
    – user1769197
    Dec 7, 2022 at 9:56
  • $\begingroup$ If two trades are run in parallel (or with overlap) for ex. buy Stock A and Stock B in the morning, sell them in the afternoon then the combined return is avg(rt). Only if the trades are sequential can you use the profits from the first trade in the second in which case the combined return is cumprod(1+rt) - 1. So the precise timing, and the existence of overlap is a crucial issue. $\endgroup$
    – nbbo2
    Dec 15, 2022 at 9:45
  • $\begingroup$ in my case, I am only trading one security, so buy and selling it in non-overlapping time interval. In that case , i can use cumprod(1+rt) - 1 then $\endgroup$
    – user1769197
    Dec 16, 2022 at 1:55

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.