Trailing Stop — Protecting Profits With a Moving Stop (ATR, Chandelier)

Anyone who has held a position for more than a day knows this pain: you were up 18%, didn't close, and rode it back to zero. Or the opposite: you closed at +5%, and the market went on without you to +60%. Both pains are — partially — cured by the same tool: the trailing stop, a stop loss that follows price as the market moves in your favor and freezes when the market reverses. Paper profit gradually turns into locked-in profit, and the position stays alive as long as the trend does.

Sounds like a free lunch — and that's where the "no hype" part begins. A trailing stop doesn't add edge; it trades one kind of cost for another: you buy profit protection with earlier exits from moves that would otherwise have kept running. Whether that trade is worth it depends on the strategy — and it can be measured.

What a Trailing Stop Is

The mechanics for a long position: the stop hangs at a fixed distance below price. Price rises → the stop rises with it. Price falls → the stop stays put (never, under any pretext, does it move down — that's an iron rule). Eventually the market reverses enough to touch the stop, and you exit — giving back the last piece of the move but keeping the rest.

The whole art comes down to one question: how do you measure the distance? Three main schools:

1. Percentage — a stop, say, 8% below the price high. Simple, but blind: 8% on BTC in a calm week is a canyon, in a week of panic it's the width of a single candle. The same number sometimes chokes a position, sometimes doesn't protect it at all.
2. ATR trailing — a stop at a distance of k×ATR below price (or below the extreme of the move). ATR measures current volatility, so the stop widens on its own when the market goes wild and tightens when the market calms down. This solves the main flaw of the percentage stop, which is why ATR is the standard. Typical multipliers: 1.5-2.5× on calm markets and shorter plays, 3-4× (sometimes up to 6×) on choppy markets and when hunting long trends. An important technical detail: a candle's ATR is only known once it closes — a touch of the line mid-candle isn't yet a signal, unless you deliberately decide otherwise and stay consistent about it.
3. Chandelier Exit — the best-known ATR trailing formula (Chuck LeBeau): stop = highest high since entry − 3×ATR(22). Anchoring to the extreme of the move (rather than to the current price) means the stop rises in steps behind the highs and doesn't react nervously to individual candles.
4. Structural (a bonus, outside the world of indicators) — a stop below successive higher swing lows, or below an N-day low. This is how the exit in the turtle trading system works: a break of the 20-day low ends the position. Advantage: the exit has a market interpretation ("trend broken"), not just a geometric one.

What the Numbers Say

QuantifiedStrategies tested ATR trailing stops as an exit mechanism in strategies on US indices; the general takeaways from their research on stops (data from QS summaries — the source file wasn't available in our research, so verify specific figures on the source page before citing them) form a consistent and inconvenient picture:

• In trend-following strategies, a trailing stop makes structural sense: the whole approach's outcome depends on a handful of very long moves, and trailing is the only mechanism that lets you ride them "to the end" — a fixed take profit cuts off the right tail of the distribution this style lives on. Empirically, systems in this family (Donchian/turtles, trend following) have used trailing exits for decades.
• In mean-reversion strategies, the test results are reversed: adding stops — including trailing ones — usually hurts results. A mean reversion move by nature dips first and recovers later; a stop that cuts the position at the bottom systematically converts future gains into realized losses. It's one of the more provocative, repeatable results in QS research.
• Tight vs. wide trailing is a pure trade-off, not a free-lunch optimization: a tight multiplier raises the share of positions closed in profit but cuts the average gain (kicking you out of trends too early); a wide one captures whole trends but gives back a deep chunk of paper profit on every reversal and deepens individual losses.

The overarching conclusion: the exit mechanism has to match the entry logic. The question "does a trailing stop work" is framed wrong — the right question is: "in MY strategy, does trading earlier exits for profit protection have a positive payoff?" Only a backtest on your rules and your market can answer that.

How to Apply It Step by Step

A practical framework for a long position on crypto (D1), Chandelier variant:

1. Entry and initial stop: enter according to your strategy (e.g., a breakout from consolidation); place the initial stop classically — behind structure, as in the stop-loss guide. Trailing usually doesn't take over right away — the position has to earn its protection first.
2. Trailing activation: after a move of ~2×ATR in your favor (or once you hit 1R), switch from static management to Chandelier: stop = highest high since entry − 3×ATR(22).
3. Update: after each daily candle closes, recalculate the level. The stop only rises — if the formula comes out lower than the current stop, do nothing.
4. Exit: either a daily close below the line (the conservative variant) or an intraday touch (the mechanical variant — the stop order sits live on the exchange). Pick one and don't waver: mixing variants under emotion is a back door to "surely it'll bounce back."
5. Position size: calculated from the initial stop using the %-risk model — trailing manages the profit, it doesn't take you off the hook for the risk math at entry.

Numerical example on BTC (illustrative): long entry after a breakout at 90,000, initial stop at 86,000, ATR(22) = 2,500. The market rises; at a price of 96,000 you activate the Chandelier: peak of the move 96,500 − 3×2,500 = 89,000 — the stop is still below entry, the trend has to work a bit more. Two weeks later the peak of the move is 108,000, and ATR has risen to 3,200: stop = 108,000 − 9,600 = 98,400 — from this point the position has a locked-in minimum of +$8,400 per BTC. The market peaks at 112,000 and reverses; the daily candle closes at 97,900, below the line → exit. You captured about +$8,900 out of a $22,000 maximum paper profit from entry to peak. Does it sting? It's supposed to — that's the price for the fact that if, instead of a reversal, price had continued to 150,000, you'd still be in the position. A fixed TP at +10% would have closed you at 99,000 and the dilemma wouldn't exist — but neither would the shot at the long tail.

When It Doesn't Work and the Most Common Traps

• Mean-reversion strategies. Worth repeating, because it's the most common structural mistake: in mean-reversion approaches, trailing (and stops in general) usually hurts results in testing. You bought a dip expecting a bounce; a stop that cuts the bottom turns the strategy's own mechanics into a machine for realizing losses.
• Trailing from minute one. A tight trailing stop activated immediately after entry is the best way to let ordinary noise eject you from a good position before the move has even started. Use a structural stop first; activate trailing after 1R-2×ATR of profit.
• A sideways market. In consolidation, trailing generates a series of "noise" exits — it protects gains that don't exist. Trailing is a trend tool; in a range, a fixed target on the opposite edge works better.
• Crypto volatility and wicks. On BTC/altcoins, a single wick can knock out a stop sitting live on the exchange and be back in the trend fifteen minutes later. Protection: a wider multiplier (3-4×ATR instead of 2×), a stop on candle close instead of intraday — and the awareness that each of these protections has its own price in the scenario of a real crash.
• Moving the stop down. "I'll give it a bit more room" — the moment the trailing stop stops existing and hope starts existing instead. A trailing stop moves in one direction. Always.
• Optimizing the multiplier to three decimal places. If a backtest shows that 2.7×ATR makes money while 2.5× and 3.0× both lose, you haven't found a parameter — you've found overfitting. Look for broad plateaus where neighboring values give similar results.

The no-hype conclusion: a trailing stop is an honest tool that does exactly one thing — it automates the decision "how long do I ride the profit" and takes it off the shoulders of emotion. It doesn't raise your win rate, doesn't create edge, and always sends a bill: every locked-in profit costs you some trend you didn't fully ride, and every fully ridden trend costs you a piece of the top you gave back. In trend-following strategies that bill historically balances out; in mean reversion it doesn't. Before you set any multiplier, answer a question more important than the parameter itself: does your strategy live off a long tail of gains, or off a high win rate? Because a trailing stop only serves the former.

FAQ

Rafał — Krypto Bez Ściemy

Trader and founder of Strefa Tradingu. He’s been breaking crypto down on YouTube for years — no hype, no signals, with a focus on market structure and risk management.

kryptobezsciemy.plInstagramYouTubeKlub Zero