Almost every trader learns to use a Stop Loss long before they learn where to put one. The usual first answer — "20 pips" or "$50" — quietly reverses the logic of the whole trade: it decides how much you're willing to lose first, then places the stop wherever that amount happens to land on the chart, regardless of whether that price level means anything. The result is a stop that gets hit by perfectly ordinary movement, on a trade whose idea was never actually invalidated. The correct order is the opposite one. Find the price at which the reason you took the trade stops being true, put the stop past it, and then calculate position size from that distance.
The Stop Marks Where the Idea Is Wrong
Every trade rests on a premise. "This pair is in an uptrend and bouncing off support." "This range has held four times and will hold again." "This breakout is real." Each of those premises has a price at which it is demonstrably false — a level where you would look at the chart and no longer want the position, regardless of what you paid for it.
That price is where the stop belongs. If you bought a bounce off support, the premise dies when price closes convincingly below that support; if you sold a rejection from resistance, it dies when price accepts above it. A stop placed there does something a fixed-pip stop can't: it makes exits automatic and unemotional, because the decision was made while you were still thinking clearly. Once the stop is set, the loss is capped and known, and every remaining decision is about managing the winner rather than negotiating with the loser — which is most of what separates a plan from the common mistakes that ruin accounts.
Structure-Based Stops
The most common professional approach places the stop beyond the nearest meaningful swing high or swing low. For a long, that means below the swing low the entry bounced from; for a short, above the swing high it was rejected at. The logic is direct: in an uptrend, price is supposed to make higher lows, so a trade to the downside through the last low is the market telling you the structure has changed.
Two refinements matter in practice. First, place the stop past the level, not exactly on it — a stop sitting at the exact price of an obvious swing low is sitting where a great many other stops sit, and price routinely trades a few pips through such levels before turning. Second, use the wick, not the body: the low of the candle's shadow is the real extreme the market rejected, and a stop between the body and the wick is inside the noise by construction. The same support and resistance levels you use to enter a trade define where its stop belongs.
Volatility-Based Stops with ATR
The structural approach answers "where is the idea wrong?" but not "how much room does this pair normally need?" ATR answers the second question directly by measuring the average distance a pair travels per bar, and it's the reason every Expert Advisor on this site sizes its stop as a multiple of ATR rather than in fixed pips.
The mechanics are simple: read the current ATR on your trading timeframe and place the stop 1.5× to 2× that distance from entry. If ATR on the H1 chart is 12 pips, a 2× stop sits 24 pips away — far enough that a single ordinary bar won't reach it. The same 24 pips on a pair with an ATR of 40 would be shredded within the hour. This is also why a stop distance that "worked last month" often stops working: volatility changed, and the fixed number didn't.
The strongest version combines both methods rather than choosing between them. Find the structural level first, then add an ATR buffer beyond it — a stop at swing low minus 1×ATR respects the chart and the pair's current volatility at once, which is exactly what the diagram above shows.
What Not to Base a Stop On
Two anchors feel intuitive and are consistently destructive. The first is the account: "I'm willing to lose $50, so the stop goes 50 dollars away." The market has no idea what your account balance is, and a level chosen from it has no reason to hold. The fix is not to abandon the $50 limit — it's to apply it in the right place, by choosing the stop from the chart and then shrinking the position size until the distance to that stop equals $50. That is the entire method taught in Risk Management Basics, and Lot Sizes Explained covers the arithmetic.
The second is round numbers and broker minimums. A stop at exactly 1.1000, or exactly 10 pips because that's the tightest your broker allows, is a stop placed for reasons that have nothing to do with your trade. If the structurally correct stop is wider than you're comfortable with, the answer is a smaller position — or skipping the trade — not a closer stop.
Practical Placement Rules
- Never widen a stop once the trade is live. Moving it further away converts a defined, planned loss into an open-ended one, and it is almost always driven by not wanting to be wrong. Tightening it, or trailing it behind structure as the trade moves in your favour, is a different thing entirely and is fine.
- Account for the spread. A long is closed at the Bid, so a wide spread effectively pulls your stop closer than it looks on the chart. On pairs with wide spreads, add the spread to the buffer.
- Watch the swap on stops held overnight. A stop that sits just beyond a level can be reached by accumulated swap and rollover costs on a position held for weeks.
- Set it when you enter, not after. A stop you intend to add "once the trade settles" is not a stop; the moment you most need it is the moment you'll least want to place it.
- Use a hard stop, not a mental one. A stop order sitting on the broker's server executes whether or not you're at the screen, and whether or not you've talked yourself into one more candle.
Worked Example
You want to buy EUR/USD at 1.0850. The most recent swing low is 1.0820, and the H1 ATR is 15 pips. Structure says the idea fails below 1.0820; a 1× ATR buffer puts the stop at 1.0805 — a distance of 45 pips from entry. Your account is $5,000 and you risk 1%, so $50 is at stake.
On a standard lot, one pip of EUR/USD is worth about $10, so 45 pips risks $450 — nine times too much. Divide: $50 ÷ (45 pips × $10 per pip per lot) = 0.11 lots. That's the position. Note what did not happen: the stop never moved to accommodate the size. With a 2× Take Profit at 1.0940, the trade risks $50 to make $100, and the only variable that flexed was the number of lots.
A Word of Caution
A correctly placed stop bounds your loss under normal conditions, not all conditions. During a major news release or a weekend gap, price can jump straight past your level and fill at the next available price — the slippage that turns a planned 45-pip loss into a larger one. Guaranteed stops exist at some brokers for a fee; the more common defence is simply not carrying large positions into scheduled high-impact events on the economic calendar. Equally, no stop placement rule improves a bad entry. If a trade needs a stop so wide that a sensible position size becomes negligible, the honest read is usually that the entry is in the wrong place, not that the stop is.