The only question that sets the level
Before any method, one question decides everything: at what price would I no longer believe this trade?
Not "how much am I willing to lose" — that is the risk amount, and it belongs to position sizing. Not "where is a round number". The stop marks the price at which the reason you entered has stopped being true. If you bought because a level held, the stop goes below that level. If you bought because a trend was intact, it goes where the trend would be broken.
Get that right and the rest is arithmetic. Get it wrong and no amount of sizing discipline compensates, because you will be stopped out of trades that were working.
The four methods, ranked
| Method | How it works | Verdict |
|---|---|---|
| Structure | Below the swing low, the range boundary, the moving average that has been holding. | Best. The level means something to other participants, so violating it is real information. |
| Volatility (ATR) | A multiple of Average True Range below entry. | Strong. Adapts to the instrument and to conditions. Best when structure is unclear. |
| Time | Exit if the trade has not worked within N bars or days. | Useful as a supplement. Frees capital from trades going nowhere; poor as the only stop. |
| Percentage | A fixed 2% (or similar) below entry. | Weakest. Arbitrary. 2% is noise on one instrument and a catastrophe on another. |
The percentage stop is popular because it is easy and feels systematic. It is not: the same 2% is a routine intraday wiggle in a high-beta technology name and a major break in a utility. It encodes no information about the instrument you are actually trading.
Volatility stops, and why ATR is the usual choice
Average True Range measures how far an instrument typically moves in a period, including gaps. A stop set at a multiple of ATR scales automatically: wider on volatile instruments, tighter on calm ones, and wider on everything when conditions become turbulent.
stop = entry − (ATR × multiple)
Common multiples run from 1.5× to 3× ATR. The trade-off is the whole decision:
- Tighter (1.5×) — smaller loss per trade and larger position size for the same risk, at the cost of being stopped out more often by ordinary noise.
- Wider (3×) — fewer false exits, smaller position size, and a larger move required before the trade pays.
Neither is correct in the abstract. What matters is that the choice is deliberate and consistent, and that you know which one your strategy's historical results were produced with. Changing the multiple mid-strategy invalidates everything you learned from the previous trades.
Worked example: the same trade, four stops
Long a $58.20 stock, $40,000 account, 1% risk ($400)
Recent swing low $55.90. Daily ATR $1.40.
| Method | Stop | Distance | Shares | Exposure |
|---|---|---|---|---|
| Structure (below swing low) | $55.80 | $2.40 | 166 | $9,661 |
| ATR 1.5× | $56.10 | $2.10 | 190 | $11,058 |
| ATR 3× | $54.00 | $4.20 | 95 | $5,529 |
| Percentage (2%) | $57.04 | $1.16 | 344 | $20,021 |
Every row risks the same $400. Look at the last one. The 2% stop sits $1.16 below entry — less than one ATR. This stock routinely moves $1.40 in a day, so that stop will be hit by a completely ordinary session, over and over, on trades that would have worked. And to keep the risk at $400 it demands 344 shares: half the account committed as exposure, on a stop that noise will reach by lunchtime.
That is the percentage stop's real failure. It is not just arbitrary — it systematically produces the largest positions behind the flimsiest levels.
The mistake that looks like discipline
You want a bigger position. The sizing formula will not give you one at your current stop. So you move the stop closer, the arithmetic obliges, and the risk figure still reads 1%.
Nothing has technically been violated, and that is exactly why this is so common. But you have replaced a stop placed where the idea would be wrong with a stop placed where your preferred size required it — which guarantees it will be hit by noise on trades that were fine.
Stop first, on the chart's merits. Size second, from the stop. Never the other way round. If the resulting size is too small to be worth the effort, the honest answer is that this trade does not suit this account — and that is allowed to be the answer.
Trailing stops and when they help
A trailing stop moves in your favour and never against you, converting an open profit into a floor. It answers a different question from the initial stop: not "was I wrong?" but "how much of this gain am I willing to give back?"
The honest trade-off is that trailing stops reduce your average winner in exchange for protecting more of them. A trend-following strategy whose profitability depends on a few very large winners is often actively harmed by a tight trail, because it exits the exact trades that were supposed to pay for everything else. A mean-reversion strategy, where gains are modest and giving them back is the main risk, benefits.
Two rules make them work. Set the trail distance on the same basis as the initial stop — usually an ATR multiple — so it reflects the instrument rather than a round number. And decide the rule before entry, because a trail invented while sitting on an open profit is not a system, it is a nervous impulse.
What a stop cannot do for you
- It is not a guarantee of price. A standard stop becomes a market order when triggered, and fills at whatever is available. In a gap or a fast market that can be far from your level.
- It does not survive gaps. Overnight news, earnings and weekend events all reopen the market away from your stop. This is the main reason to watch notional exposure as well as the risk figure.
- A stop-limit protects price but not the exit. Adding a limit stops you filling at a terrible price, and can leave you holding a losing position that runs straight through your level. Choose deliberately which of the two risks you prefer.
- It cannot rescue a bad thesis. A stop caps a single loss; it does not make a negative-expectancy strategy profitable. It buys time to find out whether the edge is real.
Once the level is set, position sizing turns it into a size, and risk of ruin covers how much to risk in the first place.