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.

Share of gross edge left after execution cost Bar chart showing that a 1.5 pip round-trip cost removes 30 percent of a 5 pip target but only 3 percent of a 50 pip target. 0%25%50%75%100% 70%5 pip target85%10 pip target93%20 pip target97%50 pip target edge kept after a 1.5 pip round trip
Share of gross edge surviving a 1.5 pip round trip. Grey is the gross target, green is what reaches the account.

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:

cost = spread + slippage + commission

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:

net edge = target − costtarget

the share of gross edge that survives to the account

Worked example at a 1.5 pip round-trip cost. Costs vary by pair, broker and session; the point is the shape, not the constant.
TargetCostEdge keptCost over 500 trades
5 pips1.570%750 pips
10 pips1.585%750 pips
20 pips1.592.5%750 pips
50 pips1.597%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.

Backtests are where this hides. A backtest that fills at the mid price and ignores slippage will show an edge that the same strategy cannot reproduce live, and the shorter the target, the larger the gap.

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.

Dilinx runs an AI trading agent with its full record published on Myfxbook, refreshed here every night.

Request Onboarding

© 2026 Dilinx Inc. All rights reserved. This article and its figures are the property of Dilinx Inc. and may not be reproduced or redistributed without written permission. Nothing here is investment advice. Trading involves risk and past performance does not guarantee future results.