Order size changes the fill: test with one lot versus ten
Posted: Tue Nov 03, 2026 5:07 am
Order size changes the fill: test with one lot versus ten
A scalper who tests a strategy on 0.1 lots and then scales to 5 lots is often shocked by how different the results feel. The strategy did not change, but the execution did.
Small orders typically fill from the top of the book at the quoted price. Larger orders chew through multiple levels. Some brokers also route small and large tickets differently, with bigger tickets getting extra checks or a longer hold time.
The practical test is straightforward. Run the same strategy during the same session with several sizes, perhaps 0.1, 1, and 5 lots, and compare average slippage and fill latency across them. The relationship is rarely linear. There is often a threshold, something like three lots, past which cost climbs quickly.
Knowing that threshold gives a real capacity limit for the system. Beyond it, either split orders across time, spread across multiple accounts, or accept lower returns per trade.
Better to find the ceiling in testing than during a live drawdown.
Where did the cost curve bend sharply on your account, and at what size?
A scalper who tests a strategy on 0.1 lots and then scales to 5 lots is often shocked by how different the results feel. The strategy did not change, but the execution did.
Small orders typically fill from the top of the book at the quoted price. Larger orders chew through multiple levels. Some brokers also route small and large tickets differently, with bigger tickets getting extra checks or a longer hold time.
The practical test is straightforward. Run the same strategy during the same session with several sizes, perhaps 0.1, 1, and 5 lots, and compare average slippage and fill latency across them. The relationship is rarely linear. There is often a threshold, something like three lots, past which cost climbs quickly.
Knowing that threshold gives a real capacity limit for the system. Beyond it, either split orders across time, spread across multiple accounts, or accept lower returns per trade.
Better to find the ceiling in testing than during a live drawdown.
Where did the cost curve bend sharply on your account, and at what size?