Trailing Stop Loss Explained: Pick the Right Trail for Your Trade

Risk warning. Leveraged products carry substantial risk to your capital. A protective exit reduces exposure; it never removes it, because gaps, slippage, and execution failures all sit outside your control.

The Direct Answer

A trailing stop is a protective exit that follows price in the profitable direction and never moves backward. Start a long trade, watch price rise, and the exit level climbs behind it. Price falls back to that level, the trade closes, and whatever gain accumulated between entry and exit is realized.

Three trailing methods dominate, and they differ in what sets the distance:

Method	How the exit moves	Main strength	Main weakness
Fixed trail	A set number of pips, points, or currency behind the best price reached	Simple, consistent, easy to automate	Ignores whether markets are calm or wild
ATR trail	A multiple of the true range behind the best price	Adapts as volatility changes	Results depend heavily on the period and multiplier chosen
Structural trail	Beyond the most recent confirmed swing, plus a buffer	Follows what the chart is actually doing	Subjective unless swing rules are defined precisely

Which suits you depends on whether you value simplicity, volatility adaptation, or alignment with chart structure. There isn’t a universally correct answer, and anyone claiming otherwise is selling something.

The 2020 version of this article covered only the third method, through one Bitcoin example, without formalizing the rule. All three are documented below, and that Bitcoin trade now appears where it belongs: as a labelled historical case study of the structural method.

Static Exits Against Moving Ones

Worth establishing the distinction before anything else.

A static protective level sits where you placed it at entry and stays there. It answers one question: how much am I prepared to lose on this idea?

A trail answers a different question: how much of an accumulated gain am I prepared to give back before leaving?

Aspect	Static exit	Moving exit
Purpose	Caps the initial downside	Protects unrealized gains
Movement	None	One direction only, never backward
When it applies	From entry onward	Usually after some threshold of progress
What it costs you	Nothing beyond the planned loss	Some portion of the best price reached
Emotional effect	Requires acceptance upfront	Requires accepting incomplete exits

Notice the second row. A trail that can move backward is not a trail, it’s an excuse, and traders who widen exits as price approaches them are the reason most accounts fail. The one-direction rule is what makes this whole technique work.

Notice also that these are complementary rather than alternative. Every trade needs a static stop at entry. The trail comes afterward, once the trade has moved far enough that protecting progress makes sense.

Fixed-Distance Trails

The simplest version, and where most platforms start you.

You choose a constant distance. The exit level follows the best price achieved since entry, staying that distance behind, and it advances only when a new best price prints.

Long example. Enter at 1.1000 with a 50-pip trail. Price reaches 1.1060, and the exit moves to 1.1010. Price continues to 1.1100, and the exit follows to 1.1050. Price then falls back, hits 1.1050, and the trade closes with 50 pips banked.

Short example. Enter at 1.3000 with the same distance. Price drops to 1.2900, and the exit descends to 1.2950. Every new low pulls it further down. A rally back to 1.2950 ends the trade.

Settings to define before running this:

• Activation point: Does the trail start immediately, or only once the trade shows a defined gain? Starting immediately on a tight distance produces exits on ordinary noise.
• Update frequency: Tick by tick, or only at candle close? Tick-based reacts faster and exits more often on spikes.
• Units: Pips, points, currency amount, or a percentage of price. Points and pips differ by a factor of ten on five-digit quotes.
• Break-even trigger: Many people move to breakeven first, then begin trailing afterward.
• Who manages it: Your platform, your broker’s server, or code you wrote. This matters more than it sounds, and there’s a section on it below.

Where a fixed trail works well: rule-based systems needing consistency across every trade, instruments with stable behaviour, and anyone who wants something they can describe in one sentence.

Where it struggles: volatility changes. A distance suited to quiet conditions gets triggered constantly when ranges expand, while the same distance in a fast market barely protects anything. The method has no way of knowing which environment it’s operating in.

What MetaTrader Actually Gives You

The built-in trailing stop deserves a walkthrough, since it’s what most readers will reach for first, and it has one property nobody mentions.

Right-click an open trade in the terminal window, choose Trailing Stop, and pick a distance from the list: 15 points, 20, 25, 30, or a custom figure. The platform then manages the exit for you.

Here’s a demonstration from a live account, which I ran on GBPAUD purely to show the mechanism. Entry at 1.8091. Price moved up, I attached a 2-pip trail, and the level followed just behind. Price ticked to roughly 1.8098, then turned, and the trade closed at 1.8096 for a gain slightly under four dollars.

Two pips is absurdly tight, obviously. I used it because I wanted the demonstration to complete inside a few minutes rather than a few hours. Any real application would use considerably more room.

The property nobody mentions: MetaTrader manages this client-side. Your terminal calculates the level and sends modification instructions to the broker. Close the platform, lose your connection, or let the machine sleep, and the trail stops advancing. Whatever level was last sent remains, but nothing moves it further.

That single fact explains why serious automated trading happens on a virtual private server. It also explains why broker-side alternatives, where they exist, behave differently.

ATR-Based Trails

The Average True Range indicator measures how far an instrument typically moves across a given period, which makes it a natural way to size a trail relative to current conditions.

The calculation:

• Long positions: exit level = highest price since entry − (multiplier × ATR)
• Short positions: exit level = lowest price since entry + (multiplier × ATR)

Worked example on a Bitcoin short:

Input	Value
Current ATR	$120
Multiplier	2
Lowest price since entry	$9,700
Resulting stop level	$9,940

If price drops to $9,500 while ATR stays at $120, the level descends to $9,740. If volatility expands and ATR reaches $200, the same multiplier produces a $400 distance, giving the trade more room precisely when markets are moving more.

That adaptation is the whole appeal. A fixed distance treats a calm Tuesday and a central bank announcement identically. This doesn’t.

Parameters requiring decisions:

Parameter	Consideration
ATR period	14 is conventional; shorter reacts faster, longer smooths
Multiplier	Typically between 1.5 and 3.5; determines how much room
Timeframe	ATR on H1 differs enormously from ATR on D1
Update timing	On completed candles, or continuously
Backward movement	Should always be prohibited
Initial exit	Whether your first protective level also uses the true range

On multiplier selection: Lower values exit sooner and bank less of the available move. Higher values survive ordinary pullbacks and return more unrealized gain before triggering. Somewhere around 2 to 3 suits trend-following approaches in my experience, though that’s a preference rather than a finding, and I’ve never tested it systematically enough to claim otherwise.

Where an ATR trail works well: instruments whose volatility varies substantially, longer-horizon positions, and anyone wanting one setting that behaves sensibly across different market conditions.

Where it struggles: the readings lag. Volatility that collapses suddenly leaves the distance too wide for a while, and volatility that explodes leaves it too narrow until the average catches up. Also, two people using different periods and multipliers are running genuinely different systems while both calling it the same thing.

Structure-Based Trails

Here the exit level references what the chart has done rather than a calculated distance.

Formal rules, which the original article never stated:

Direction	Rule
Long	After price prints a new higher high, move the exit below the most recent confirmed higher low, plus a buffer
Short	After price prints a new lower low, move the exit above the most recent confirmed lower high, plus a buffer

Both rules contain two conditions that must be met in sequence. The swing must be confirmed, and price must then extend beyond the previous extreme. Moving the exit as soon as a swing appears, before the extension confirms it, is the most common way this method goes wrong.

Confirming a swing needs its own definition, or the whole thing collapses into judgment:

• A fractal, meaning a high with two lower highs on each side.
• A high with three or more lower highs on each side, filtering more noise.
• A ZigZag reversal exceeding a set percentage.
• A swing spanning at least some multiple of recent volatility.

Pick one. Write it down. Apply it identically every time, including when the chart looks tempting and your rule says wait.

Where the structural method works well: trending markets, price-action approaches, and traders who want their exits to sit where other participants’ orders plausibly cluster.

Where it struggles: swings form slowly. Between one confirmed swing and the next, a trade can give back a substantial portion of its progress while the rule waits for confirmation. In fast reversals, the level you’re protecting can be a long way behind current price.

Choosing a Buffer

Placing the exit exactly at a swing extreme invites trouble, since price frequently probes just beyond a level before continuing. A buffer sits between the structural point and your exit.

Buffer method	Example	Notes
Fixed distance	20 points beyond the swing	Simplest, ignores conditions
Volatility-based	0.25 × ATR beyond the swing	Adapts, needs a period defined
Spread multiple	Two times the current spread	Scales with execution costs
Percentage	0.1% beyond the level	Works across instruments with different price scales

Decide the method before you open the trade. Choosing it while watching an open trade is how discretion sneaks back into what should be a rule.

Round numbers deserve a mention here. Price often reacts around them, and orders cluster there, so a buffer placed just beyond a round figure can make sense. That reasoning is also easily abused as post-hoc justification for whatever level feels comfortable, which is precisely the error I made in the case study below.

Comparing the Three Methods

Feature	Fixed	ATR-based	Structural
Adapts to volatility	No	Directly	Indirectly, through swing spacing
Objectivity	High	High	Medium unless swings are defined
Automation difficulty	Low	Low	Moderate, needs swing logic coded
Early exit risk	Higher when markets expand	Lower with a suitable multiplier	Depends on swing frequency
Gain returned before exit	Predictable	Expands and contracts	Can be substantial between swings
Best suited to	Simple rule-based systems	Volatility-sensitive approaches	Trend-following and price-action methods
Setup complexity	Minimal	Two parameters	Swing rule plus buffer rule

My own preference leans toward the second for automated systems and the third for anything I’m watching manually, though I’d hold that loosely. Plenty of profitable traders use nothing but a fixed distance and never think about it again, which has an appeal I sometimes envy.

Break-Even Moves Are Not Trails

Frequently conflated, genuinely different.

Moving to break-even means shifting the protective level to your entry price once the trade shows some progress. It happens once. Afterward the level sits still.

A trail adjusts continuously as price advances.

Many people combine them: move to break-even at some threshold, then let the trail take over beyond it. Reasonable, though be aware that break-even exits are not actually free. Costs still apply, so an exit at your exact entry produces a small loss after spread and commission. Placing it slightly beyond entry handles that.

Partial Exits and Multiple Positions

Closing part of a trade at a target while trailing the remainder is common, and it changes the arithmetic in ways worth understanding.

• Banking part of the gain reduces the emotional weight of the rest, which improves adherence to rules.
• The remaining portion can run considerably further than a fixed target would have allowed.
• Average outcome per trade typically falls while variance falls further.
• Whether this improves overall results depends on how often your instrument produces extended moves.

Managing several positions simultaneously adds complexity. Each needs its own level, and platforms differ in whether a trail applies per trade or per instrument. Assign distinct identifiers when running automated management across multiple positions, or you’ll find one adjustment overwriting another.

The Exposure You Started With

A trail manages progress. It does nothing about how much you committed at the outset, and the original article never addressed this.

Before entry, decide:

1. Maximum acceptable loss on the idea, as a percentage of the account.
2. Initial protective level, placed where the reasoning behind the trade fails rather than where the loss feels tolerable.
3. Volume, calculated from the distance between entry and that level.
4. Whether gaps can jump past it, which they can, particularly on cryptocurrency and over weekends.
5. Total exposure across everything currently open.

Trailing without this groundwork protects gains on positions that were sized wrongly from the start.

Manual, Platform, Broker, or Code?

Four ways to run any of these methods, with meaningfully different failure modes.

Approach	How it works	What breaks it
Manual adjustment	You move the level yourself	Sleep, travel, distraction, hesitation
Platform-managed	The terminal calculates and sends modifications	Closing the terminal, lost connection, machine sleeping
Broker-managed	The server holds the instruction	Availability varies; not every broker offers it
Custom code	An Expert Advisor manages it	Coding errors, platform restarts, unhandled edge cases

The client-side limitation covered earlier applies to the second row and catches people out constantly. A trail that stops advancing when your laptop sleeps is not the protection you thought you had.

Broker-managed versions solve that, since the instruction lives on the server, but availability is inconsistent and the distance options are often limited to what the firm offers.

Custom code gives complete control over the stop and introduces its own problems. Points to handle explicitly when writing it:

• Prohibit backward movement under all conditions.
• Update on completed candles rather than every tick, unless you specifically want tick-level reaction.
• Check the broker’s minimum distance requirement before sending modifications, since exits placed too close to current price are rejected.
• Handle multiple positions with distinct identifiers.
• Log every modification, so you can reconstruct what happened afterward.
• Decide behaviour when the platform restarts mid-trade.

Hosting on a virtual private server removes the sleep and connectivity problems from your trading setup for both platform-managed and coded approaches. Many brokers provide it free once account balance passes some threshold.

Indicator-Driven Trails

Beyond the three main methods, some traders let a technical indicator carry the stop level directly.

Common variants include a moving average, where the level sits at the average value and rises with it, or a Parabolic SAR, which was designed for precisely this job and accelerates as a move extends. Chandelier exits combine the highest price reached with a volatility multiple, which is really an ATR trail wearing a different name.

Indicator	How it sets the level	Behaviour
Moving average	Level tracks the average value	Smooth, lags badly on sharp reversals
Parabolic SAR	Accelerates toward price over time	Tightens automatically as the move matures
Chandelier	Best price minus a volatility multiple	Close cousin of a standard volatility trail
Donchian channel	Opposite side of an N-period range	Structural in effect, mechanical in calculation

Advantage: these are unambiguous, since an indicator produces one number and there is nothing to interpret. A stop driven this way automates cleanly and remove the swing-confirmation problem entirely.

Disadvantage: the level goes wherever the calculation puts it, which occasionally means somewhere the chart offers no support at all. A moving average that has drifted far from price during a strong run leaves an enormous amount of unrealized gain exposed.

My own trading leans toward volatility-based distances for automated systems and structural levels when watching manually, though several traders I respect use nothing but a Parabolic SAR and never think about the question again.

Where Trails Fit in a Trading Plan

Something the original article never addressed: none of this belongs in isolation.

Any trading plan defines what you enter, how much you commit, where the idea fails, and how you leave. Trails answer only the last part, and only for trades that have already moved favourably. Roughly half of everything you open will never reach the point where any of this matters, which is worth remembering before spending a weekend optimizing multiplier values.

Where trails genuinely change outcomes is on the minority of trades that run. Trend-following approaches derive most of their results from a small number of extended moves, and how you exit those determines whether the approach works at all. Mean-reversion approaches, by contrast, typically target a defined move and rarely benefit much from letting winners run, since the whole premise is that price returns to some average rather than continuing indefinitely.

So match the exit mechanism to the entry logic. A trail bolted onto a mean-reversion system usually just returns gains that a fixed target would have banked cleanly.

What Traders Get Wrong About Volatility

Worth separating two ideas people frequently merge.

Volatility affects how far price travels in a given period. It does not tell you direction, and a wide trailing distance during turbulent conditions is not a prediction that the move continues; it’s an acknowledgement that ordinary noise now spans more ground.

Traders often tighten distances when markets become volatile, reasoning that they want to protect gains before things move against them. That reaction is understandable and usually counterproductive, because the tighter level sits inside the range of ordinary movement and closes the trade on noise rather than on a genuine reversal.

The volatility-adjusted approach does the opposite by construction, widening exactly when most people want to tighten. Whether that feels comfortable is a separate question from whether it works, and I’d suggest testing it on a demo account before deciding based on instinct.

A Practical Setup Checklist

Before running any trail on real capital:

1. Decide which of the three methods matches your entry logic.
2. Define every parameter in writing, including activation threshold and update timing.
3. Confirm your broker’s minimum distance requirement for the instrument.
4. Establish whether management is client-side, server-side, or coded.
5. Arrange hosting if execution must continue while you are away.
6. Test across at least a few hundred historical trades using tick-level modelling.
7. Compare results against a fixed target on identical entries.
8. Run it on a demo account through both calm and turbulent conditions.
9. Record every exit and whether the level behaved as intended.
10. Only then apply it to a live account, at reduced size.

Steps six and seven get skipped most often in retail trading, and they’re the two that tell you whether the mechanism helps your particular approach or merely feels sophisticated. Most people adopt a moving stop because it sounds like disciplined risk control, then never check whether it improved anything. Sometimes it doesn’t, and finding that out on historical data costs nothing beyond an afternoon.

Common Mistakes

• Widening the level as price approaches: Converts a planned exit into an unplanned larger loss.
• Setting the distance too tightly: Ordinary noise closes positions that would have worked.
• Starting immediately at entry: Before a trade has any progress, a tight trail is just a narrow static exit.
• Choosing the buffer while watching: Decide beforehand.
• Assuming the platform keeps working when you close it: A client-side stop is only as reliable as your connection.
• Ignoring gap risk: Exits are instructions, not guarantees, and price can open past them.
• Applying one distance across every instrument: Fifty pips means very different things on EURUSD and on a yen cross.
• Never testing the rule: Which brings us to the next section.

Testing a Trail Rule Properly

Adding a trail changes a strategy’s entire outcome distribution, so it needs testing rather than assuming.

• Compare against a fixed target: Run identical entries with a fixed exit and with each trail method, then compare average result, win rate, and largest decline. Frequently the trail lowers the win rate while raising the average outcome, and whether that’s an improvement depends on your temperament as much as the arithmetic.
• Test parameter sensitivity: If a 2.0 multiplier works and 1.8 or 2.2 collapse, you found noise rather than structure. Robust settings degrade gently.
• Watch the modelling mode: Trail logic depends on the sequence of movement inside each candle, which coarse simulation modes never reconstruct. Tick-level testing matters more here than for entry logic.
• Include costs: Every modification is free, but the exit itself pays spread and possibly slippage, and trails produce more exits from adverse moves than fixed targets do.
• Forward test before trusting it: Simulated exit fills are optimistic. Real ones, particularly during fast reversals, land where they land.

Frequently Asked Questions

Does a moving stop guarantee I keep my gains? 

No, and treating it as a guarantee causes real damage. The instruction closes your trade when price reaches the level, assuming a counterparty exists at that moment and the market is open. Weekend gaps, news gaps, and sudden illiquidity can all produce fills well beyond the intended point. Cryptocurrency and index products gap more readily than major currency pairs. Treat the level as your intended exit rather than a guaranteed one, and size positions accordingly.

Which trailing method suits beginners best? 

The fixed version, almost always, because it has one parameter and produces behaviour you can predict before it happens. Learning what a moving stop does to your trading results matters more initially than optimizing the method, and complexity added before you understand the basic mechanism tends to obscure what’s actually driving outcomes. Move to volatility-based or structural approaches once you’ve watched a fixed distance behave across a few hundred positions.

Can I combine two methods? 

Yes, and hybrid arrangements are common. A frequent pattern uses a volatility-based distance normally, with an override that jumps the level whenever a confirmed swing offers a closer point. Another moves to break-even at a fixed threshold, then switches to trailing beyond it. Complexity carries a cost, though: every additional condition makes the rule harder to test and easier to rationalize adjusting mid-trade. Two mechanisms is usually plenty.

What ATR period should I use? 

Fourteen is the conventional default and a reasonable starting point, since it balances responsiveness against stability across most instruments. Shorter periods react faster to changing conditions and produce more variable distances, while longer ones smooth out short-term swings at the cost of adapting slowly. The timeframe you calculate on matters more than the period itself: a reading taken from hourly candles produces a completely different scale from one taken daily.

Do trailing stops work on stocks and indices? 

The mechanics transfer to any instrument, though gap behaviour differs substantially. Shares gap overnight on earnings and news, and indices gap over weekends, so an exit intended at one level can fill considerably beyond it. Distances also need rethinking, since a percentage that suits a currency pair may be far too tight on a volatile share. Test each market separately rather than carrying settings across from something else.

Should the level move on every tick or only at candle close? 

Candle-close updating produces fewer adjustments and avoids reacting to brief spikes that reverse immediately, which suits most approaches holding positions for hours or longer. Tick-based updating captures more of a favourable move before reversing, at the cost of exiting on noise. Shorter timeframes generally favour tick updates while longer horizons favour candle close. Whichever you pick, test both, because the difference in results is often larger than expected.

What happens if my exit sits too close to current price? 

Most brokers enforce a minimum distance, and modification requests inside it get rejected. Your platform typically logs an error while the previous level stays where it was, meaning you believe you’ve adjusted something that never changed. Check the minimum distance in your instrument’s contract specification, and build the check into any automated management you write. This is among the most common silent failures in coded exit logic.

Is a trail better than a fixed profit target? 

Neither is universally better, and they suit different market behaviour. Fixed targets perform well when moves tend to travel a predictable distance then reverse, producing a higher win rate and consistent outcomes. Trails perform well when occasional moves run far beyond typical range, capturing extended runs a target would have cut short. Look at your instrument’s historical behaviour: if large trends are rare, targets probably suit you better.

How much of my gain will a trail typically return? 

Roughly the trailing distance itself, since the trade closes that far from the best price reached. A 50-pip trail returns approximately 50 pips of whatever the maximum unrealized gain was, plus slippage. That arithmetic explains why distance selection is a direct trade-off: tighter keeps more of small moves while exiting large ones prematurely, and wider survives volatility while returning more at the end.

Can I test a stop rule on a demo account first? 

You should, and for longer than feels necessary. Watching a moving exit behave across different conditions teaches you more than any explanation, including this one, particularly regarding how often it closes positions you would have kept manually. Demo fills are optimistic, so treat the results as a best case. What transfers reliably is your understanding of the mechanism and whether you can tolerate exits that feel premature.

Disclosure: Educational content only, not a recommendation to trade. Broker links elsewhere on this site may carry commercial arrangements. Historical examples describe past conditions and do not indicate future behaviour. Consider guidance from a regulated professional before committing capital.