ATR-based sizing is common, but there's a refinement worth considering: comparing today's session volatility to its own history and cutting size when it's at an extreme.
How it works:
1. Calculate the 15-minute ATR at, say, 08:15 UK each day.
2. Keep the last 100 values.
3. Find where today's value sits as a percentile.
Then a rule:
- Below 20th percentile: normal size, but expect smaller targets.
- 20th–80th: normal size.
- 80th–95th: 75% size.
- Above 95th: 50% size or no trading.
Why reduce size in high volatility if your stop is already scaled with ATR? Because extreme days bring other risks that ATR doesn't capture: slippage, spread spikes, gaps through stops and emotional pressure. A 12-pip stop on a crazy day fills worse than a 6-pip stop on a normal day.
Low-volatility days have their own problem: the cost becomes a larger share of the expected move. Some traders reduce trade count instead of size in that case.
Do you use any percentile-based rule like this, or do you stick to straight ATR scaling?
Sizing down when session volatility is in an extreme percentile
-
FXS London Scalper
- Posts: 806
- Joined: Sun Oct 18, 2026 12:25 pm
Sizing down when session volatility is in an extreme percentile
FXS Official Team | London Session Desk