Do stop-losses actually work? What they protect and what they cost
The short answer
Stop-losses work as advertised for traders — they enforce an exit you'd otherwise talk yourself out of. For long-term investors they usually hurt, because a percentage-based stop is triggered by volatility rather than by anything about the business, and normal market noise will repeatedly sell your good positions near local lows. The useful question isn't whether to use one, it's whether your exit rule is tied to price or to your thesis.
What a stop-loss actually is
A stop-loss is a resting instruction: if the price hits X, turn my position into a market order and sell.
Two things follow immediately, and both get glossed over.
It's a trigger, not a floor. Once it fires you sell at whatever the market offers next, which in a fast decline can be materially below your stop. A stop-limit avoids the bad fill but introduces the opposite failure — in a gap-down, the limit never fills and you're still holding, with the added false comfort of having "had protection."
It knows nothing about the company. A 15% stop fires identically for a stock that dropped because earnings collapsed and one that dropped because the whole market had a bad week. The instruction is about price only.
Where they genuinely work
Stop-losses aren't a bad tool. They're a tool for a specific job, and that job is trading.
Short holding periods. If your thesis plays out over days or weeks, price is most of your information. There's no long-run business case to fall back on, so a price-based exit is coherent.
Enforcing a plan you'd otherwise abandon. The strongest argument for stops has nothing to do with markets and everything to do with people. Almost everyone holds losers too long, hoping. A stop makes the decision in advance, when you're calm, and executes it without asking you again. That's the same logic as a written falsifying condition — precommitment beats willpower — just implemented at the broker instead of on paper.
Position-level risk budgeting. If you size positions so that a stop being hit costs you a fixed, small percentage of the account, you've built a system where no single trade can seriously damage you. That's real risk management, and the stop is the mechanism that makes the arithmetic hold.
Leverage and concentration. If you're using margin or holding a very large single position, an automatic exit isn't optional — the downside is genuinely unbounded relative to your equity.
Where they quietly cost you
Now the long-term investor's case, which is the opposite.
Volatility triggers them, not risk. Individual stocks routinely draw down 20–30% in a year that ends up positive. Set a 15% stop on a normal equity position and you're not protecting against loss — you're guaranteeing that ordinary noise ejects you, repeatedly, at prices near local lows. You've converted survivable volatility into realized losses.
They sell the recovery. A stop fires into weakness by construction, which means it systematically exits near lows and leaves you in cash for the bounce. Then you face the second, harder decision: when to get back in. That's the same two-decision problem that makes market timing fail, except now it runs automatically and you didn't even choose the moment.
They fail exactly when you need them. Stops are sold as crash insurance. In a real crash — a gap down on news, a fast cascade — market stops fill well below the trigger and limit stops don't fill at all. The protection is most reliable in the mild declines you didn't need protecting from.
They can substitute for thinking. "I've got a stop on it" is a comfortable thing to say and it's often standing in for having no thesis. A stop is not a substitute for knowing why you own something; it's a rule about price attached to a position you may not be able to explain.
Taxes and costs. In a taxable account, each triggered stop is a realized event and a round trip in spread and commission. Repeated small losses compound in the wrong direction.
The better question
Instead of "what percentage should my stop be," ask: what would tell me this investment is wrong?
For a trade, price answers that — the setup failed. Use a stop.
For an investment, price is a poor proxy. The honest answers are business facts: growth fell below the level your case required, a key customer left, margins compressed structurally, management you were backing walked out. Those are the exits worth pre-committing to, and none of them are expressible as "down 15%."
This is why the four reasons to sell don't include price. A decline tells you the price changed; it doesn't tell you the business did. A stop-loss can only ever observe the first one.
There is one legitimate hybrid: a very wide stop — well beyond normal volatility for that name — used as a catastrophe backstop rather than a risk-management tool. It won't fire on ordinary drawdowns, and it's there for the case where something breaks faster than you can react. Just be honest that it's a circuit breaker, not a strategy.
Trailing stops, briefly
A trailing stop follows the price up and only sells on a retracement from the high. It sounds like it solves the problem: you keep the upside, you cut the downside.
In practice it inherits the same flaw. It's still triggered by volatility, so it systematically sells your best positions during their normal pullbacks — and your best positions tend to be the volatile ones. The classic outcome is being trailed out of a multi-year compounder at a 20% retracement in year two, which costs vastly more than the drawdown it avoided.
Trailing stops are reasonable for locking in a completed trade. They're a poor fit for a position you intended to hold for years.
Test it before you trust it
Everything above is general. What matters is what a given rule does to your portfolio, and that's testable rather than arguable.
Two things worth checking before committing to a stop policy:
- How often would it have fired on positions that recovered? Run your actual holdings against the drawdowns they've already had. Most people are surprised how many of their winners spent time down 25%.
- Do you need it as a behavioral crutch? If you have a documented history of holding losers far too long, a stop may be worth its costs — you're paying volatility tax to fix a bigger behavioral leak. That's a legitimate trade, but it should be a diagnosis, not a default.
The second question is the one that actually decides it, and it's answerable. Sydnical grades holding behavior directly — whether you cut winners early, hold broken theses, or repeat the same exit mistake across tickers — using real positions and live prices with nothing at risk. If the pattern is there, you'll see it. If it isn't, a stop is mostly a cost. The historical crash scenarios are the stress test for the same question: a stop policy that looks sensible in a calm market behaves very differently across a real one.
FAQ
Do stop-losses actually work?
They work as designed — enforcing an exit at a preset price — which is genuinely useful for short-term trades, leveraged positions, and investors who reliably hold losers too long. They work poorly for long-term investing, because they're triggered by price volatility rather than by anything about the business, so normal fluctuations sell good positions near local lows.
What is a good stop-loss percentage?
There isn't a universal number, and that's a hint the framing is off. Individual stocks commonly draw down 20–30% within years that finish positive, so any stop tight enough to feel protective will fire on ordinary noise. If you use one for long-term holdings at all, set it far outside normal volatility for that name and treat it as a catastrophe backstop.
Should long-term investors use stop-losses?
Usually not. A long-term thesis rests on business fundamentals, so the exit condition should be a fundamental change — growth falling below a threshold, losing a major customer, structural margin compression — rather than a price level. The exception is a documented tendency to hold broken positions indefinitely, where the stop's costs may be worth fixing the bigger problem.
Do stop-losses protect you in a market crash?
Less than advertised. A stop-loss converts to a market order when triggered, so in a fast decline or an overnight gap it can fill well below your intended price; a stop-limit avoids the bad fill but may not execute at all. Stops are most reliable in the modest declines where protection matters least.
What's the difference between a stop-loss and a trailing stop?
A stop-loss sits at a fixed price; a trailing stop follows the price upward and triggers on a set retracement from the high. The trailing version locks in gains but has the same underlying flaw — it's driven by volatility, so it tends to exit your most volatile positions, which are often your best long-term holdings, during normal pullbacks.
An exit rule tied to price protects the trade. An exit rule tied to your thesis protects the investment. Find out which one you're actually running →