A trailing stop is a stop-loss that moves in the direction of a profitable trade and never moves back. It turns an open gain into a floor under the position, so a retracement decides the exit instead of a target chosen in advance.
also: trailing stop-loss, dynamic stop, trail
Updated · Reviewed
A stop that follows price up but never follows it back down. It gives a winning trade room to run, and sets the exit at a level the trade has already passed. Like any stop it is only a trigger, so a gap can still close it worse.
A fixed target has to be chosen before anyone knows how far the move will go. A trail replaces that guess with a rule: stay in until the market takes back a defined amount. What that rule is worth depends on how uneven a strategy's winners are.
| Setting | Typical value | Note |
|---|---|---|
| Trail distance (ATR) | 2–3 × ATR(14) | Below roughly 1.5 × ATR the trail sits inside ordinary retracement and closes trends early. |
| Activation threshold | after 1 × the stop distance in profit | Stops the trail pulling to break-even before the trade has done anything. |
| Step (minimum improvement) | 5–10 pips | Without a step the EA modifies on nearly every tick, which brokers throttle. |
| Update frequency | once per closed bar | Tick-by-tick trailing multiplies request volume and rarely improves outcomes. |
A long entered at 1.0800 with a 20-pip trail. The stop follows the high-water mark and never retreats, so the retracement picks the exit.
A fixed 40-pip target would have finished at 1.0840. Different bets on how far moves run.
Calculation 1.0880 − 20 pips = 1.0860
Result 60 pips kept, 20 given back — the cost of not predicting the top
A trail sits between two costs. Too tight and trends end early; too loose and much of the move goes back. The distance is the whole decision.
| Range | What it means |
|---|---|
| Trail under about 1.5 × ATR | Ordinary retracement closes the position, so the strategy never sees the large moves. |
| Trail at 2–3 × ATR | The usual working range for trend strategies: outside the noise, close enough to protect most of a move. |
| Trail wider than the move so far | The stop is still behind entry, so the whole open profit can be handed back before anything triggers. |
| Trailing from the first tick, no threshold | Drags the stop to break-even at once and turns normal fluctuation into scratch trades. |
Whether a listed EA trails its exits or takes a fixed target is stated in its strategy description, and the published ledger shows what the exits realised. Every measured figure on this page comes from those ledgers (EXP-EXIT-SHAPE-001), closed-trade basis.
A trail inside ordinary retracement turns a trend strategy into a scalper wearing a trend strategy's entry rules. The win rate rises, the average win collapses, and for a strategy that depends on its largest winners the expectancy usually goes with it.
It is a trigger price like any other stop. A weekend or news gap through the level closes the position at the next available price.
The 24 records measured here keep between 1.27% and 98.16% of gross profit when their winners are capped at their own median winner. One rule, opposite consequences.
That distance is fitted to one history's retracements. A value surviving across neighbouring settings is worth more than the peak of a narrow spike.
MetaTrader’s own trailing stop is a terminal feature, not an order the broker holds for you.
Step 4 is the failure that looks like success. The position still carries a stop, the chart still shows a line, and nothing reports an error. An EA’s own trailing code has the same dependency.
Each listing carries its closed-trade list, so the distance of every exit and the money it produced can both be read directly — 17,357 trades in all.
| Measure (24 records, 17,357 closed trades) | Median | Range or count |
|---|---|---|
| Exits closing against entry: distances inside ±2% of that symbol’s median | 83.78% | above 50% on 17 of 24 |
| Exits closing in favour: the same band | 26.30% | above 50% on 7 of 24 |
| Largest tenth of winners’ share of gross profit | 25.64% | 11.85% to 34.08% |
| Largest single winner’s share of gross profit | 1.17% | up to 4.89% |
| Gross profit kept if winners are capped at the median winner | 62.83% | 1.27% to 98.16% |
| The same, capped at twice the median winner | 84.76% | — |
The two distance rows are not symmetric, and the asymmetry runs one way. No record repeats its winning distance without also repeating its losing one, while ten do the reverse: a fixed stop with a winning side that lands anywhere. The tail is also thinner than trend folklore suggests. The largest single winner reaches at most 4.89% of a record’s gross profit, so the gross profit here is spread across many winners rather than resting on one.
That shape can be checked against what each listing says it does. Six records declare a trailing exit in their published strategy description, and all six land in the dispersed group — their winning distances cluster a median 2.06%, the highest is 38.35%, and none reaches 50%. The other eighteen sit at a median of 35.03%, seven of them above 50%.
Measure the unevenness of the winners first, then pick a distance. Average true range supplies the scale, volatility says whether that scale still matches the one the settings were tuned in, and parameter optimisation covers the neighbouring-values check a trail distance has to survive.