Trailing Stop

A trailing stop rides behind the trade at a fixed distance. On a long with a $800 trail: price rises to $65,000 and the stop rises to $64,200; price rises to $66,000 and the stop follows to $65,200. If price falls, the stop stays put. It only ever ratchets in your favor:

long trailing stop = max(previous stop, highest price since entry − trail distance)

What it buys you

A trailing stop converts an open-ended winner into a defined giveback. You never exit at the top, and you also never round-trip a large gain back to zero. It's the standard exit for trend-following systems, where a few big moves carry the whole distribution of wins.

Choosing the distance

Fixed dollar or percent trails get shaken out when volatility expands, so the usual refinement is scaling the trail to ATR, commonly 2 to 3×, which adapts it to current conditions. Too tight and normal noise stops you out of good trends; too wide and you give back most of each move.

AlphaProve signals support trailing_stop_distance, and the engine ratchets the level bar by bar exactly as above.

On AlphaProve

To scale the trail to volatility as the section above suggests, the atr_trail_distance adaptive helper returns an ATR-based distance you pass straight into the signal, so the trail widens and tightens with conditions rather than staying fixed. A companion control, breakeven_at_profit_pct, jumps the stop to entry once the trade is a set percentage in profit, clearing the loss before the trail even engages. The MAE/MFE scatter on the tearsheet then shows how much open profit each winner gave back, the number that tells you a trail is too wide.