Sizing for worst-case fills instead of planned stops
Posted: Sat Oct 31, 2026 8:06 pm
Position size is usually computed from the stop distance. Risk 1 percent, stop 5 pips, calculate lots. It's neat, and it assumes the stop will fill at 5 pips.
Scalpers know the reality. Average fills may be close, but the tail includes fills at 8, 12 or 20 pips in bad conditions. Over enough trades, those tail events decide the outcome.
A more robust method sizes for a fatter loss. If typical stop slippage is 0.3 pips but the 99th percentile is 3 pips, plugging in the worse figure reduces size slightly and protects against surprises.
It's a small price. A 10 percent smaller position is not going to change annual returns much, yet it can prevent a rare disaster from becoming a serious drawdown.
The calculation also changes with the time of day. During news windows, the stress figure should be larger.
The needed data comes from the journal, which is another reason to log slippage on every stop.
Have you ever sized a trade based on the worst fill rather than the plan?
Scalpers know the reality. Average fills may be close, but the tail includes fills at 8, 12 or 20 pips in bad conditions. Over enough trades, those tail events decide the outcome.
A more robust method sizes for a fatter loss. If typical stop slippage is 0.3 pips but the 99th percentile is 3 pips, plugging in the worse figure reduces size slightly and protects against surprises.
It's a small price. A 10 percent smaller position is not going to change annual returns much, yet it can prevent a rare disaster from becoming a serious drawdown.
The calculation also changes with the time of day. During news windows, the stress figure should be larger.
The needed data comes from the journal, which is another reason to log slippage on every stop.
Have you ever sized a trade based on the worst fill rather than the plan?