Execution
Execution cost is not a rounding error
A one and a half pip round trip removes 30 percent of a five pip target and 3 percent of a fifty pip one. The same cost, the same broker, two completely different businesses.
Every trade pays a toll before it can be right. The toll has three parts and they are all charged whether the trade works or not:
measured in pips, charged on the round trip
Spread is quoted and visible. Commission is contractual. Slippage is the one that gets ignored, because it only shows up as the difference between the price you expected and the price you got, and that difference has to be recorded trade by trade or it disappears into the noise.
Cost is fixed, edge is not
The toll does not care how far you were planning to run. It is roughly constant per trade, so its weight depends entirely on target size:
the share of gross edge that survives to the account
| Target | Cost | Edge kept | Cost over 500 trades |
|---|---|---|---|
| 5 pips | 1.5 | 70% | 750 pips |
| 10 pips | 1.5 | 85% | 750 pips |
| 20 pips | 1.5 | 92.5% | 750 pips |
| 50 pips | 1.5 | 97% | 750 pips |
The last column is the same in every row. Five hundred trades cost 750 pips regardless of what you were aiming at. Against a 5 pip target that is 2,500 pips of gross edge required just to break even. Against a 50 pip target it is 25,000, but you only needed 750 of it.
What this rules out
A strategy targeting a handful of pips has to be right far more often than its longer-horizon equivalent, purely to cover the toll. That is achievable, but it moves the problem from prediction to infrastructure: fill quality, latency and spread selection start to matter more than the entry logic.
The measurement worth keeping is simple. Record the price you intended and the price you got, on every fill, and take the average. That number belongs in the expectancy calculation from the start, not as an adjustment at the end.