Why Fixed-Spread Backtests Flatter Scalping Results

[Shadow Trader]

Scalping strategies target small moves, often a few pips, so the spread and commission represent a large fraction of every trade. A backtest that assumes a constant spread of one pip can show a smooth equity curve that disappears once real costs are applied.

Real spreads are not constant. They widen around rollover, during news releases, and in thin liquidity such as the Asian session open for some pairs. A strategy that fires frequently near those windows tends to pay the most exactly when the tester charges the least.

Variable spread data, or at minimum a session-based spread table taken from the actual broker, gives a more honest picture. Commission per lot and swap should also be added, along with an allowance for slippage on market orders and stop exits, which usually fill worse than quoted.

A useful stress test is to inflate the modelled cost by fifty percent and rerun. If the net result collapses, the edge is thinner than the headline figure suggested. Strategies that survive that test have a margin of safety against the normal drift in broker conditions over time.