Chandelier Exit: Hung From the High

Michael Whitman · Trading since 2017, full-time since 2020 · Reviewed 2026-09-06

The chandelier exit places a trailing stop three average true ranges below the highest high of the last twenty-two bars, so it hangs from the top of the move rather than from the entry price. It is a rules-based exit and it produces no entry signal at all.

How it works

A candlestick chart of the site's shared price history. The headline on the chart reads: A stop hung from the highest high since entry. — A stop hung from the highest high since entry. Illustrative chart - not real market data.

The chandelier exit hangs a stop below a recent high. For a long position it sits a fixed number of average true ranges beneath the highest high of the lookback window.

A gently rising stretch of the long price series. The headline on the chart reads: Hung from the ceiling, which is where the name comes from. — Hung from the ceiling, which is where the name comes from. Illustrative chart - not real market data.

The name describes the anchoring. A chandelier hangs from the ceiling and moves up as the ceiling rises; the stop does the same thing from the highest high.

A calmly advancing stretch of the long price series. The headline on the chart reads: Highest high of twenty-two bars, minus three average ranges. — Highest high of twenty-two bars, minus three average ranges. Illustrative chart - not real market data.

The published defaults are 22 bars and 3 average ranges. Twenty-two is roughly a trading month of daily bars, and three is a deliberately loose multiple. Both are settings, not laws.

A choppy, directionless stretch of the long price series. The headline on the chart reads: It anchors to a high, not to the price you paid. — It anchors to a high, not to the price you paid. Illustrative chart - not real market data.

The anchoring is the real difference from an ordinary trail. Your entry price is an accident of when you happened to act; the highest high since then is a property of the market, and the exit reads better for it.

What the defaults deliver

A flat, quiet stretch of the long price series. The headline on the chart reads: At three ranges a trail survived a median of 22 bars here. — At three ranges a trail survived a median of 22 bars here. Illustrative chart - not real market data.

Measured on this site’s shared 576-bar history, a trail at three average ranges lasted a median of 22 bars across 562 trials, with the ninetieth percentile at 47.

A strongly rising stretch of the long price series. The headline on the chart reads: And 92% of those trails were eventually stopped. — And 92% of those trails were eventually stopped. Illustrative chart - not real market data.

92% of them ended in a stop-out. The loose multiple postpones the exit and does not avoid it, which is the honest description of every trailing rule.

A declining stretch of the long price series. The headline on the chart reads: The lookback matters as much as the multiple. — The lookback matters as much as the multiple. Illustrative chart - not real market data.

The 22-bar window is the parameter that gets ignored. A shorter lookback drops old highs sooner, so the anchor falls faster after a peak and the stop tightens; a longer one keeps a distant high alive and leaves the stop far below price for weeks. Both changes move holding time in the same way the multiple does, which is why changing them together makes any result impossible to attribute.

In practice

A candlestick chart with a volume histogram beneath it, with the volume histogram emphasised. The headline on the chart reads: It reads price only, so participation is a separate check. — It reads price only, so participation is a separate check. Illustrative chart - not real market data.

Nothing in the calculation looks at volume. A high made on heavy participation and one made on nothing anchor the stop identically.

A long-horizon candlestick view of the same price series. The headline on the chart reads: Twenty-two daily bars is roughly a trading month. — Twenty-two daily bars is roughly a trading month. Illustrative chart - not real market data.

The lookback is in bars, so its meaning changes with the chart. Twenty-two five-minute bars is under two hours, and the same setting on a daily chart is a month of trading.

A candlestick series containing several opening gaps, with the largest opening gap marked. The headline on the chart reads: A gap below it exits somewhere else entirely. — A gap below it exits somewhere else entirely. Illustrative chart - not real market data.

A gap below the level fills at the open, and because the chandelier sits far below price by design, the gaps that reach it are the large ones.

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 is an exit rule and never an entry one. — It is an exit rule and never an entry one. Illustrative chart - not real market data.

It produces no entry signal. Price crossing the line is an exit for a position you already hold and says nothing about opening one.

And every exit still costs a share of a bar. Illustrative chart - not real market data.

The loose setting is cheap on turnover. A 22-bar median holding period means far fewer round trips than a tight trail, and each one is 2% of a median bar’s range on this history.

A 72-bar candlestick section of the shared price history. The headline on the chart reads: The level is computed, not defended by anyone. — The level is computed, not defended by anyone. Illustrative chart - not real market data.

The exit price is arithmetic. No participant is defending it, which argues for placing it beyond structure rather than treating it as structure.

Position size is the other half

A three-range stop is a wide stop, and a wide stop with an unchanged position size is a large loss. The distance and the size have to move together, or the looser exit simply converts a series of small losses into a smaller series of much larger ones.

The arithmetic is straightforward: risk per trade divided by stop distance gives position size. With the distance moving as volatility changes, the size changes on every trade — which is the part most people skip, and the part that decides whether a loose exit is an improvement or an expensive habit.

What a chandelier exit is not

It is not an entry system. It only manages a position you have.

It is not fixed risk. The distance moves with volatility.

It is not anchored to you. The market’s high sets the level.

And it is not gap-proof. Nothing resting at a price is.

When it fails

A sideways, range-bound candlestick series. The headline on the chart reads: In a range the ceiling never rises and the stop never moves. — In a range the ceiling never rises and the stop never moves. Illustrative chart - not real market data.

In a range the highest high stops rising, so the stop freezes and the exit behaves like a fixed one — protecting nothing that a static level would not have protected while the position goes nowhere.

The second failure is a lookback longer than the trend. A 22-bar high from three weeks ago keeps the stop absurdly far below current price and gives back most of the move.

A third is tightening the multiple after a bad exit. The setting is fitted to the trade that just finished and shortens every future one.

A fourth is running it without adjusting size. Three average ranges of risk at an unchanged size is a much larger loss than the method intends.

And a fifth is treating 22 and 3 as correct. They are the defaults from one publication, and the survival table is how you check what your own choice buys.

The original data

On this site’s shared 576-bar history, a trail at three average true ranges opened at every eligible bar survived a median of 22 bars, a mean of 25.0 and a ninetieth percentile of 47, with 92% stopped out across 562 trials. The 14-bar average true range has a median of 0.5994, ranging from 0.2823 to 0.7954 between the tenth and ninetieth percentiles. 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 high is three ranges away. Take the exit? — The high is three ranges away. Take the exit? Illustrative chart - not real market data.

The gap between the median of 22 bars and the ninetieth percentile of 47 is where the method earns its keep. A trend-following exit only works if it occasionally holds through a long move, and that ninetieth percentile is where the profit that pays for all the ordinary exits comes from. Judge a loose exit on that tail rather than on its average — the median is where most trades finish, and the tail is why the rule exists at all.

ATR trailing stop is the general form and the parameter discussion. Trailing stop has the full survival table across six distances. And stop loss placement covers choosing between structure and calculation.

What I actually do

What I like about the chandelier is that it is anchored to the market rather than to me. My entry price is a fact about my day; the highest high since I got in is a fact about the market. Anchoring the exit to the second one removed a lot of arguing with myself.

— Michael Whitman

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