Struggling to reconcile ATR-based position sizing with hard stop-loss limits
I have been trying to transition from a fixed-percentage risk model (just risking a flat 1% of equity per trade) to using Average True Range (ATR) to determine my stop distance and position size. The logic of letting the asset's current volatility dictate the room the trade needs makes a lot of sense on paper, especially when switching between different FX pairs like $GBPJPY and $EURUSD.
Where I am getting stuck is dealing with the math during high-volatility regimes. When the ATR expands significantly, my calculated invalidation level stretches far wide. To keep my total dollar risk constant, my position size shrinks to almost nothing. Sometimes the size gets so small that it barely clears my broker's minimum lot requirements, or the spread cost ends up eating a disproportionate chunk of the potential move. Conversely, if I cap the stop-loss distance to prevent this, I feel like I am completely violating the volatility-adjusted logic and just begging to get stopped out on random noise.
For those who use volatility-based sizing, how do you handle these extreme ATR expansion periods? Do you step aside entirely when the volatility implies a stop-loss that is too wide for practical execution, or do you have a secondary filter that overrides the ATR model?