A stop-loss has one job: to take you out at the price where your trade idea is objectively wrong. Everything that goes wrong with stops — hunted, too tight, too wide, moved in a panic — comes from giving it a different job, usually "keep my loss comfortable."
The principle: invalidation, not comfort
Every structured trade has a price at which its reason no longer exists. Buying a support bounce? The idea is invalid somewhere below that support. Trading a breakout? Invalid back inside the range. The stop belongs at the invalidation point — and if the resulting risk is too large for your account, the correct fix is a smaller position, never a closer stop. A stop moved inside the invalidation level converts "I was wrong" into "I was early and got charged anyway": the market can hit it and then do exactly what you predicted.
Giving the level room
Stops placed exactly at obvious levels fail for a mechanical reason: everyone can see the same level, so the orders cluster a pip or two beyond it, and that cluster is liquidity — markets routinely pierce an obvious low, fill the stops, and reverse. Two defenses. First, place beyond the noise, not at the line: below the wick extremes of the zone, not the tidy horizontal everyone drew. Second, size the buffer by volatility: a common method sets the buffer using ATR — for example, stop = invalidation level minus 0.5–1.0× the 14-period ATR. On a pair moving 80 pips a day, a 15-pip buffer is a coin flip; the same buffer on a quiet cross may be generous. ATR keeps the stop's meaning constant as conditions change.
The placement menu
- Structure stops — beyond the swing high/low or zone that defines the trade. The default for level-based trading.
- Volatility stops — a pure ATR multiple (e.g. 2× ATR) from entry, used when the entry isn't anchored to one clean level. Honest about noise, blind to structure.
- Time stops — exit if the trade hasn't performed within N bars. Underrated: a breakout that goes nowhere for two days has usually already failed, whatever the price says.
- Hard equity stops — the account-level line (e.g. down 3% on the day, stop trading). Not a chart tool; a you tool.
Trailing: paying for protection with potential
Moving the stop to breakeven or trailing it behind swings converts open profit into locked profit — and every tightening also raises the odds of being wobbled out of a move that was still valid. There is no free setting: trail tight and you keep small wins and forfeit trends; trail loose (say, behind each new higher low, or 2–3× ATR) and you give back more per trade but stay in the runners. Decide the trailing rule when you plan the trade, because deciding it while watching open profit shrink reliably produces the worst version.
The two unforgivables
Widening a losing stop. The moment you move a stop further away to avoid taking the loss, you no longer have a stop — you have a hope with paperwork. This single habit converts small planned losses into account-relevant ones, and it is behind most blowups that "only happened once."
Trading without one because "stops get hunted." The wick-hunt problem is real and the solution is placement and sizing, not exposure to unlimited loss. In a leveraged, gapping, 24-hour market, the mental stop fails exactly when it matters: in the fast move, at 3am, mid-news. If you cannot watch the screen, the stop must be on the server.
The honest trade-off
Every stop placement is a purchase: closer stops buy bigger position size and cheaper wrong-ness, paid for with more frequent stop-outs; wider stops buy staying power, paid for in size. There is no placement that avoids both costs — there is only knowing which cost you chose, before the market presents the bill.