Adding up the full cost of a short trade usually gives a higher number than people assume. A worked example for 1 lot of EURUSD on a typical raw account, with illustrative figures:
Spread at entry: 0.2 pips ($2).
Commission round turn: $7.
Average slippage on entry and exit together: 0.1 pips ($1).
Total: $10, or 1.0 pip.
Costs often forgotten:
VPS and data subscriptions spread across all trades in the month.
Spreads during your own trading hours, which can be above the advertised average.
Orders bounced or requoted, then filled at a worse level.
What it means: if the strategy targets 4 pips with a 4 pip stop, costs are a quarter of every target. A 55% win rate on paper becomes roughly break-even after costs.
To do it with your own numbers, take one month of real fills and compute the average total cost per trade in pips. Then compare it with your average winner.
What surprised you most when you added up your costs?
The useful comparison is net expectancy, not win rate alone. With equal 4-pip wins and losses, a 55% win rate gives only 0.4 pips before costs; subtract a 1-pip all-in cost and the example turns negative. Convert spread, commission and both-side slippage into pips using the actual lot size and account currency, then compare that total with realized movement.
Keep costs separate in the log: quoted spread at entry and exit, commission charged, entry and exit slippage, and any rejects or requotes. Tag pair, session, order type and news/volatility conditions, because averages hide the windows where costs jump. Use actual fills from the account being assessed rather than the advertised minimum, and verify whether commission is round-turn or charged per side.
Then calculate net expectancy by setup and hour, and test whether filtering expensive periods still leaves enough trades. A four-pip target may be too small when routine costs approach one pip, especially if slippage has a long tail. Wider targets are not automatically better, but the strategy needs a measured edge after every cost, not just on the chart.