WhitmanTrading

Trailing Stop: What the Distance Buys

A trailing stop follows price in the favourable direction and never moves against the position. The only decision it requires is the distance, and on this site's shared history that distance bought a median of 3 bars at one average true range and 22 bars at three.

How it works

A candlestick chart of the site's shared price history. The headline on the chart reads: A stop that follows price up and never comes back down.
A stop that follows price up and never comes back down. Illustrative chart - not real market data.

A trailing stop is a stop that moves with price. As a long position gains, the stop rises behind it; when price falls, the stop stays where it is.

A gently rising stretch of the long price series. The headline on the chart reads: It moves one way, which is the entire mechanism.
It moves one way, which is the entire mechanism. Illustrative chart - not real market data.

The one-way ratchet is the whole idea. It removes the decision to take profit and replaces it with a rule, and it converts an open profit into a floor that only ever rises.

A calmly advancing stretch of the long price series. The headline on the chart reads: The distance is the only decision and it decides everything.
The distance is the only decision and it decides everything. Illustrative chart - not real market data.

And the distance is the only input. Everything about how the position behaves — how long it lasts, how much it gives back, how often it gets clipped — follows from that one number.

What each distance actually buys

A choppy, directionless stretch of the long price series. The headline on the chart reads: At one average true range it lasted a median of 3 bars.
At one average true range it lasted a median of 3 bars. Illustrative chart - not real market data.

Opened at every bar of this site’s shared 576-bar history and trailed by one average true range, the median position survived 3 bars. Across 562 trials, every single one was eventually stopped.

A flat, quiet stretch of the long price series. The headline on the chart reads: At two it lasted 10, and at three it lasted 22.
At two it lasted 10, and at three it lasted 22. Illustrative chart - not real market data.

Doubling the distance more than tripled the survival time. Two average ranges gave a median of 10 bars; three gave 22. That relationship is steeply non-linear at the tight end.

A strongly rising stretch of the long price series. The headline on the chart reads: At four it lasted 32, so the gain flattens out.
At four it lasted 32, so the gain flattens out. Illustrative chart - not real market data.

Then it flattens. Going from three to four ranges — a third wider, and a third more risk on every trade — bought 10 more bars, against the 12 that the step from two to three delivered for a smaller increase.

A declining stretch of the long price series. The headline on the chart reads: And 91% of them were stopped out eventually.
And 91% of them were stopped out eventually. Illustrative chart - not real market data.

Widening does not prevent the stop-out; it postpones it. Even at four average ranges, 91% of trails were eventually hit. A trailing stop is a device for deciding when to leave, not whether to — which reframes the distance question entirely: you are buying holding time, at a known price in risk.

In practice

A candlestick chart with a volume histogram beneath it, with the volume histogram emphasised. The headline on the chart reads: In thin conditions the same distance is hit far sooner.
In thin conditions the same distance is hit far sooner. Illustrative chart - not real market data.

A fixed distance behaves differently by session. In thin volume a wider spread and erratic prints reach the same level far sooner than in active conditions.

A long-horizon candlestick view of the same price series. The headline on the chart reads: On a daily chart those bar counts become months.
On a daily chart those bar counts become months. Illustrative chart - not real market data.

The survival numbers are in bars, so they scale with the chart. Twenty-two bars is under two hours on a five-minute chart and a month on a daily one.

A candlestick series containing several opening gaps, with the largest opening gap marked. The headline on the chart reads: A gap through it fills wherever the market reopens.
A gap through it fills wherever the market reopens. Illustrative chart - not real market data.

A gap through the trail fills at the open, which is the one case where the protection the trail appears to provide does not exist.

A declining stretch of the long price series, with the entry price and the level at which a stop would trigger drawn as horizontal lines. The headline on the chart reads: It never moves against you, which is the whole point.
It never moves against you, which is the whole point. Illustrative chart - not real market data.

Moving a trail down is not a variation, it is abandoning it. The one-way rule is the mechanism, and a trail that can retreat is an ordinary stop with extra steps.

A candlestick chart of the site's shared price history, annotated with the round-trip cost. The headline on the chart reads: Each stop-out and re-entry costs a share of a bar.
Each stop-out and re-entry costs a share of a bar. Illustrative chart - not real market data.

A tight trail plus re-entry is an expensive pattern. Each cycle is a round trip at 2% of a median bar’s range on this history, and a 3-bar median survival implies a great many cycles.

A 72-bar candlestick section of the shared price history. The headline on the chart reads: A trailing stop is a resting order that keeps moving.
A trailing stop is a resting order that keeps moving. Illustrative chart - not real market data.

Check how your broker implements it. A server-side trail rests at the venue; a platform-side one only works while the platform is running. Some trail on the last trade and some on the bid or ask, which changes where it triggers.

Choosing the distance

Work backwards from holding period rather than from comfort. If a method expects to hold for roughly ten bars, a trail that survives a median of three is going to end most positions before the method’s own thesis has played out — and that mismatch is invisible until it is measured.

Then check the arithmetic on the other side. A wider trail gives back more from the peak on every winning trade, and the survival table above says how much time each step buys. The right distance is the one where holding time matches the method’s expected duration, which is a calculation rather than a preference.

What a trailing stop is not

It is not free protection. Wider costs more per trade.

It is not a profit target. It exits behind price, never ahead.

It is not gap protection. A gap fills wherever it reopens.

And it is not a substitute for an initial stop. It has to start somewhere.

When it fails

A sideways, range-bound candlestick series. The headline on the chart reads: In a range it is hit before the move ever starts.
In a range it is hit before the move ever starts. Illustrative chart - not real market data.

A range is where it does its worst work. Price oscillates enough to trigger the trail repeatedly without ever travelling, so the method pays a full round trip for each cycle and ends the period flat in price and down in costs.

The second failure is tightening after giving profit back. The trail that would have kept the last trade is fitted to the last trade, and it shortens every future one.

A third is a distance chosen from comfort. How much you are willing to give back is a feeling; how long you need to hold is a number.

A fourth is trailing a position that was too large to begin with. No exit rule fixes a size problem.

And a fifth is expecting it to prevent losses. Between 91% and 100% of trails on this history ended in a stop-out, whatever the distance.

The original data

On this site’s shared 576-bar history, a long trail was opened at every bar with an average true range value — 562 trials at each distance. At one average range the median survival was 3 bars and 100% were stopped; at 1.5 it was 5 bars and 98%; at two, 10 bars and 96%; at 2.5, 16 bars and 93%; at three, 22 bars and 92%; at four, 32 bars and 91%. The figures are in research/series-measurements.json, produced by site/measure_series.py.

A 72-bar window of the shared price history, cut short at the decision bar. The headline on the chart reads: The trail is one bar away and the trend looks fine. Widen?
The trail is one bar away and the trend looks fine. Widen? Illustrative chart - not real market data.

The stop-out rate barely moves and the survival time triples. Between one and four average ranges the proportion eventually stopped falls only from 100% to 91%, while the median holding period goes from 3 bars to 32. So the distance is almost purely a decision about time, not about safety — and the useful way to set it is to run this same table on your own instrument, then pick the multiple whose median survival matches how long your method actually needs to be in the trade.

Stop loss is the fixed version and where the initial level comes from. ATR trailing stop sets the distance from volatility. And take profit is the alternative exit that works ahead of price.

What I actually do

I spent years tightening trails after giving profit back, and every tightening cost me more than it saved. The measurement that changed my mind was survival time - once I could see that halving the distance roughly thirds the holding period, it stopped being a judgment call.

— Michael Whitman

This page is educational, not financial advice. Test every idea on your own charts before risking money.