Chandelier Exit Explained: When the Stop Line Moves Down

A Chandelier Exit is drawn by subtracting a volatility distance from the highest high the market has reached inside a lookback window. For a long position the line sits below price, and the usual instruction is to leave the trade when price closes through it.

Described that way it sounds like a trailing stop that only tightens. The formula does not behave that way. The highest high it measures is the highest high inside a moving window, and a window moves in both directions, so the line can fall while a position is still open and the distance the trader is risking can grow without any decision having been made.

This page works through the behaviour first, then the settings, then the gap between a line on a chart and an order sitting at a broker.

The Line Is Not a Ratchet, and Where It Moves Against You

A trailing stop, in the sense most traders mean it, only moves one way. It follows price up in a long position and stays where it is when price falls back, so the worst outcome is locked in one step at a time. Every general description of the Chandelier Exit reads as if it does this.

The formula contains nothing that enforces it. The reference point is the highest high inside a window of the last so many bars, and that window slides forward with every new bar. When the bar that produced the extreme high finally drops out of the back of the window, the reference high becomes whatever the highest remaining bar reached, which is lower. The subtraction then runs from a lower number and the exit level falls.

Worked through with figures, holding the volatility distance constant at 240 pips for the illustration: the window’s highest high is 1.1000, so the line sits at 1.0760. Some bars later the 1.1000 bar has aged out of the window and the highest bar still inside it reached 1.0930. The line is now 1.0690. It has moved 70 pips further from price, in a long position, while nothing about the trade changed.

Whether that matters depends on what the line is being used for. As an exit signal read on the close of each bar it is a nuisance and no more. As the level a stop-loss order is dragged to, it means the risk on an open position quietly widened, which is the opposite of what a trailing stop is chosen to do.

What the Formula Measures, and What It Does Not

Two inputs produce the line. The first is an extreme: the highest high for a long, the lowest low for a short. The second is a volatility distance, the average true range multiplied by a number the trader chooses. The exit level is the extreme with that distance taken off it, and the short version is the same arithmetic with the sign flipped, which is why it needs no separate explanation.

What the formula does not measure is the entry. The distance is not risk per trade and it was never anchored to the price paid, so the level says nothing about whether the position is ahead or behind.

That is the cleanest way to separate it from the two volatility stops already covered on this site. How ATR is read as a stop distance covers the average true range itself and the fixed stop placed a multiple of it away from an entry price, and that page is where the indicator basics belong rather than here. The N volatility unit of the turtle rules is measured from the entry as well, and drives position size before it ever places a stop.

Why a Lookback Stops Meaning One Month Below the Daily Chart

Explanations of this indicator almost always justify the lookback by calling it a month of trading days. That reasoning holds on one timeframe and nowhere else, because the setting is a count of bars and a bar is whatever length the chart says it is. The same number that reaches back a month on a daily chart reaches back roughly a day on an hourly one and a few hours on a fifteen-minute one.

The consequence is not cosmetic. A shorter reach in real time means the extreme ages out of the window sooner, so the line drops more often and by smaller amounts. Traders who move down a timeframe without touching the setting keep a number chosen for a reason that no longer applies to what they are looking at.

The extreme itself also behaves differently. A window on an intraday chart can sit entirely inside one session, which makes the reference high a session high rather than a swing high. A highest-high channel covers what a window of extremes does and does not tell you, including why a level drawn from one broker feed will not match another exactly.

A Plotted Line Is Not a Resting Order

An indicator draws on a chart. A stop-loss order sits on a broker server at a price, and the two have no connection unless something creates one. Following a Chandelier Exit therefore means either modifying the order by hand each time a new bar closes, or running an expert advisor that reads the line and modifies it automatically.

Left alone, the order stays where it was last placed. That is fine when the line is rising, since a stale order is simply a looser one. It becomes misleading in the case described above: the line falls, the trader believes the exit has widened with it, and the resting order is still at the tighter level. The chart and the account disagree.

There is a second reason the two rarely match exactly. The line is computed from bar prices, and a stop order is filled against the market at the moment it triggers. A resting stop order and what happens when it is hit is worth reading before wiring any indicator to one.

What a Volatility Spike Does to the Stop Distance

The second input moves too, and it moves fastest when the market is least comfortable. The average true range rises after any bar with a wide range, so the distance subtracted from the extreme widens after a shock rather than before it. If the volatility distance doubles from 240 pips to 480, the exit level drops by 240 pips on its own, with the reference high unchanged.

That behaviour is defensible in one reading: a wider distance keeps the position out of noise that has genuinely grown. It is uncomfortable in another: the level loosens precisely when the market has demonstrated it can travel further in a single bar.

Both effects land at once when a market reopens away from where it closed. The gap bar produces a wide true range, and the exit level takes a step down that has nothing to do with the trend the indicator is meant to follow. A weekend gap is the common case in currencies, and the page on it covers what an order sitting inside that gap can expect.

The Averaging Method the Documentation Never Names

The volatility half of the formula depends on the average true range, and the definition of true range is not in dispute. MetaQuotes documents it as the greatest of three quantities: the current bar range, and the distance from the previous close to the current high and to the current low.

The averaging is where the precision runs out. The MetaTrader 5 documentation defines the indicator as a moving average of true range values and does not state which moving average, and the MQL5 function that returns it exposes a single averaging period and no smoothing choice.

Wilder, who introduced the average true range, used a smoothing of his own. A simple average of the same data gives a different number, and a different number multiplied by the same multiplier puts the line somewhere else.

For a reader this matters only in one practical way, and it is worth knowing before comparing charts: two tools set to identical inputs can plot this line at two different prices, and neither is wrong. The number to trust is the one from the tool the orders are actually placed through.

Who This Is Not For

This indicator assumes a trend it can follow and a position already open. In a range it produces an exit level that hovers below the range low in longs and above the range high in shorts, far enough away to be useless as a stop and too far to be informative.

It also does nothing for entries, nothing for position size, and nothing for the question of how much of an account should be at risk on one trade. Those are separate decisions and this line does not inform any of them.

Which of the three uses fits depends on how the output is consumed. Read on the close of each bar as a signal to leave, the falling-line behaviour costs nothing and the indicator does what it was built to do.

Used as the level a stop-loss order is dragged to, it needs a rule of your own that refuses to move the order down, because the formula will not supply one. Used on an intraday chart with a setting inherited from a daily one, the first thing to fix is the setting, not the indicator.