Use Mark-to-Market Returns for Sharpe Ratio Calculations
Summary
The document considers whether daily strategy returns for a Sharpe ratio should use realized profit and loss from cash alone or include the market value of open positions. The answer favors marking the account to market each day, so gains and losses accrue as the positions change in value rather than appearing as a large cash movement only when a trade closes. The example contrasts a steady sequence of daily gains with the same total gain concentrated on the final day, illustrating how realized-only accounting can distort measured return volatility.
The discussion also notes that cash interest matters, particularly when the strategy is not fully invested or uses leverage, since idle cash can earn interest and borrowed balances can incur charges. It gives practical accounting guidance but does not specify a full return methodology, treatment of fees, or valuation conventions for illiquid positions. Those details still need to be consistent in a performance evaluation.
Key ideas
- Daily performance should reflect the mark-to-market value of open positions.
- Recording profit only when a position closes can concentrate returns on exit dates and distort volatility.
- Sharpe ratio inputs should use a consistent daily account valuation.
- Cash interest and borrowing costs can affect returns when capital is idle or leverage is used.
- The discussion does not prescribe detailed valuation rules for illiquid holdings or transaction costs.
Tags
Full text
# Calculate Daily Returns for Sharpe Ratio
# Calculate Daily Returns for Sharpe Ratio
For the purposes of Sharpe ratio, I calculate a trading strategy's daily returns using realized P/L only: $$ \frac{K(t + 1) - K(t)}{K(t)}, $$ where $K(t)$ is the cash balance after market close on day $t$. Assume no transfer is made from or to the account.
Recently, someone suggested I use account balance(cash balance + market value of all positions) instead, to include the day's unrealized P/L.
Which one should I use?
## Answer by RaveTheTadpole (score 4, accepted)
https://quant.stackexchange.com/a/9941
I would absolutely use a mark-to-market value in your daily pnl for the purposes of evaluating performance (e.g. Sharpe). So, yes, that would include the value of open positions in addition to your cash balance.
If you hold something for a year, that performance was earned one day at a time, not all at once. If you only look at cash, you will have a large cash flow when you exit a position, and that will overstate the volatility of your returns. The difference in standard deviation between [0,0,0,0,0,0,0,0,99] vs [11,11,11,11,11,11,11,11,11] is significant, even though they total the same.
The effect of interest might be relevant (as BlueTrin as suggested), but is secondary to the importance of using mark-to-market.
## Answer by BlueTrin (score 2)
https://quant.stackexchange.com/a/9940
For a single day as long as $K(t+1)$ includes the intraday cash flows it is the same, however if you do not simulate your cash balance interest rate you forget that your cash get compounded over time, which is slightly incorrect.
This is why someone suggested you simulate your cash balance. This is more correct as well if you are not always 100% invested or if you have access to leverage as you get interest or get charged for being in credit/debit.Shown in full with attribution under the source's licence. Licence: CC BY-SA 4.0 (Stack Exchange)
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.