Chandelier Exit

An ATR-based trailing stop that hangs a fixed volatility distance below the highest high (or above the lowest low) and only ever ratchets in the trade's favor.

What it is

The Chandelier Exit is a volatility-based trailing-stop method attributed to Chuck LeBeau and popularized in Alexander Elder's writing. The name describes the mechanism: the stop 'hangs' from the ceiling of the recent price range — for a long position, a multiple of the Average True Range (typically 3 × ATR) subtracted from the highest high of the lookback window (typically 22 bars); for a short position, the mirror image above the lowest low. The critical detail most explanations omit is the ratchet: the stop is only allowed to move in the trade's favor. Without that condition, a volatility spike or a new range would let the stop retreat, which defeats the purpose of trailing. It is important to be clear about what this tool is: a method for placing and moving an exit level, not an entry signal and not a source of edge. A stop method can control how a loss is taken; it cannot guarantee a profit or improve the odds of any individual trade. This is educational material only, not financial advice, and all trading carries risk of loss.

How it works

Three ingredients drive the calculation. First, ATR over the lookback (default 22 bars) measures recent typical bar-to-bar range, including gaps. Second, the extreme of the window: ta.highest(high, 22) for the long-side stop, ta.lowest(low, 22) for the short side. Third, the multiplier (default 3.0) sets how many ATRs of adverse movement are tolerated. The raw long stop each bar is highest-high minus 3 × ATR. The ratchet then compares this to the stop's own previous value: if the prior close was above the previous stop (the long context is intact), the stop becomes the maximum of the raw value and the previous stop — so it can rise or hold, never fall. If price closes below the stop, the ratchet resets and the short-side logic (lowest low + 3 × ATR, ratcheting downward) takes over. Because ATR adapts to volatility, the stop trails loosely in volatile conditions and tightens in quiet ones — by design, it gives a position more room exactly when bars are wider. Note what the level is not: it is derived purely from past highs, lows, and ranges. It knows nothing about where price will go, and a gap can trade straight through it, so the plotted line is a decision level, not a guaranteed fill price.

How traders read it

• Long-side line below price: the level at which, by this method's rules, a long position would be considered stopped — the distance to it is always roughly 3 ATRs from the recent high, i.e. a volatility-scaled cushion.
• Ratcheting: the long stop rising bar after bar simply reflects new highs or falling volatility; a flat stop means neither has happened. The stop never moving down is the defining property — if a plotted version does move down mid-trend, the implementation is missing the ratchet.
• Crossings: price closing through the stop flips the active side (long stop below price becomes short stop above). Frequent flips indicate a ranging market where any flip-stop method churns.
• Distance from price: a wide gap between price and the stop means high recent ATR or a strong run from the extreme; it also means a larger adverse move is tolerated before exit.
• Multiplier intuition: 3 × ATR roughly means 'exit after an adverse move three times the recent average bar range' — reading it as a volatility budget is more useful than treating the line as support or resistance.
• Remember the line only describes past range mathematics; gaps and fast markets can move through it without filling at that price.

Common settings

The classic parameters are a 22-bar lookback (about one trading month on daily charts) for both the ATR and the highest-high/lowest-low window, with a multiplier of 3.0. Smaller multipliers (1.5-2.5) trail tighter, exiting sooner on both noise and genuine reversals; larger ones (3.5-4+) tolerate deeper pullbacks in exchange for giving back more when a move ends. Some implementations use close instead of high/low for the extreme, which produces a tighter line. There is no correct setting — the trade-off between exiting on noise and giving back open gains is unavoidable, only shiftable.

Strengths

• Volatility-adaptive: the exit distance is measured in ATR units, so it automatically widens in volatile regimes and tightens in quiet ones rather than using a fixed price or percentage offset.
• The ratchet enforces discipline mechanically: the stop only moves in the trade's favor, removing the temptation to loosen a stop mid-trade.
• Fully deterministic and non-repainting when computed from confirmed bars — the level for a given bar never changes after that bar closes.
• Anchoring to the highest high (rather than to price or an average) means the stop trails the best point of the move, a genuinely different behaviour from EMA- or price-offset trails.
• Simple, auditable math built from two standard components (ATR and rolling extremes).

Pitfalls to watch

• Missing ratchet in many ports: naive implementations recompute highest-high minus 3 × ATR each bar, letting the stop drop after volatility spikes — behaviourally a different (and less coherent) tool.
• Whipsaw in ranges: as a flip-stop, it alternates sides repeatedly in sideways markets; the method assumes a trend to trail and has nothing useful to say without one.
• Late by construction: giving a 3-ATR cushion means roughly 3 ATRs of open gain are surrendered whenever the move ends — trailing stops trade earlier exits for bigger giveback, always.
• Gap risk: the stop is a level, not a fill. Price can gap far beyond it, and real exits in fast markets can be materially worse than the plotted line.
• Settings sensitivity: lookback and multiplier changes move every exit; a combination that looked clean historically is a description of that history, not a property of the market.
• No edge, no prediction: a stop method manages how positions end. It cannot make a losing approach profitable, and no exit rule removes the risk of loss.

Pine v6 example

//@version=6
indicator("Chandelier Exit Example", overlay = true)

len  = input.int(22, "ATR / Lookback Length", minval = 1)
mult = input.float(3.0, "ATR Multiplier", minval = 0.1, step = 0.1)

atrDist = mult * ta.atr(len)

// Long stop: hangs below the highest high, ratchets up only
longStop = ta.highest(high, len) - atrDist
longStop := close[1] > nz(longStop[1], longStop) ? math.max(longStop, nz(longStop[1], longStop)) : longStop

// Short stop: hangs above the lowest low, ratchets down only
shortStop = ta.lowest(low, len) + atrDist
shortStop := close[1] < nz(shortStop[1], shortStop) ? math.min(shortStop, nz(shortStop[1], shortStop)) : shortStop

plot(longStop,  "Chandelier Long Stop",  color = color.teal)
plot(shortStop, "Chandelier Short Stop", color = color.maroon)

stopHitLong = ta.crossunder(close, longStop) and barstate.isconfirmed
plotshape(stopHitLong, "Long stop crossed", style = shape.xcross, location = location.belowbar, color = color.orange, size = size.tiny)

Pro tip: When auditing a Chandelier Exit script, watch the long-stop line through a volatility spike inside an uptrend: a correct implementation holds or rises (the ratchet), while a broken one dips as ATR expands. That single visual check catches the most common porting error. And treat the line as a pre-planned decision level, not a promise of execution — gaps trade through stops. Educational only, not financial advice; no exit method guarantees outcomes, and all trading involves risk of loss.

Educational only — not financial advice, not a recommendation to trade. No indicator is predictive; trading involves substantial risk of loss.