Stop-losses: the exit you choose in advance

A stop-loss is the only trading decision you get to make while you are still calm. Every other choice — whether to hold, whether to add, whether this dip is the bottom — gets made later, by a version of you that is watching money disappear and negotiating with reality. The stop is a message from the rational you to the future you, and it says: if price reaches here, the idea was wrong, get out.

That framing matters more than any placement technique. A stop is not a prediction of where price will bounce. It is the price at which your reason for being in the trade no longer exists. Everything in this lesson follows from that definition.

What a stop actually does — and does not

Mechanically, a stop-loss order rests dormant until price touches your trigger level, then fires an order to close the position. On Obsidiate, the stop-loss order in the trade terminal executes as a market order once triggered, which means it prioritizes getting you out over getting you a precise price. In normal conditions the fill lands near the trigger. In fast or thin conditions it can land meaningfully worse — a stop guarantees an exit attempt, not an exit price. We will deal with that honestly in the lesson on gaps; for now, hold both truths: a stop is the most reliable risk tool you have, and it is not a force field.

Placement by structure

Structure-based placement puts the stop where the market itself would prove you wrong. If you bought because a level held — say a stock bounced three times at $48 and you entered at $50 — then the trade thesis dies if price closes below $48. A stop at $47.60, just beyond the level, is placed at the point of invalidation. The logic is sound because it ties your exit to evidence rather than to your comfort.

• Place stops beyond the structure, not at it — a level that held three times will be probed to the tick.
• Wider structure means wider stop means smaller position, never the reverse. Size adjusts to the stop; the stop never adjusts to the size you wanted.
• If the nearest sensible structure is so far away that the position would be tiny, that is information: the entry is bad, not the stop.
• Round numbers ($50, $1.00, $20,000) attract clustered orders; placing your stop a little beyond them, not exactly on them, avoids the crowd.

Placement by percentage

Percentage-based placement ignores the chart and exits at a fixed loss — 2%, 5%, 8% below entry. Its strength is discipline: it caps loss mechanically and cannot be rationalized in the moment. Its weakness is that the market does not know your percentages. A 5% stop on an instrument that routinely wanders 6% in a quiet session is a donation schedule. The percentage approach works best when the percentage is calibrated to the instrument: tighter on a calm FX major, wider on a small-cap coin whose daily range would eat a stock trader alive. Many experienced traders blend the two methods — find the structural level first, then check that the implied loss as a percentage of the account is acceptable, and skip the trade if it is not.

Stop hunting, discussed like adults

Stop hunting is real, but mostly not in the cinematic sense. The popular version — a shadowy market maker watching your individual $400 stop and driving the price to it — overestimates everyone’s interest in you. The realistic version is structural: stops cluster at obvious places (just below support, under round numbers, beneath the prior day’s low), and clustered stops are a pool of guaranteed liquidity. Large traders who need to buy size benefit from pushing price into that pool, because triggered sell-stops become market sells they can buy from. No conspiracy required; just incentives.

• The defense is not removing your stop — that converts a small planned loss into an unplanned large one.
• Place stops where the thesis dies, not where everyone else’s pain begins; the two are usually a few ticks apart.
• Expect levels to be overshot before they hold. Budget for the wick.
• If you keep getting stopped at the exact low, your stops are too obvious and too tight — widen the placement and cut the size to keep risk constant.

Mental stops versus real stops

A mental stop is a promise to exit manually when price reaches a level. In theory it offers flexibility and protection from wicks. In practice it has a failure rate that should disqualify it for most people, because it asks for perfect discipline at the precise moment discipline is hardest. The sequence is depressingly standard: price hits the mental stop, the trader decides to give it a little room, the loss doubles, and now exiting means admitting twice the mistake — so they hold. Markets do not punish the indecision politely. A real, working stop order has none of these failure modes. It does not feel hope. It is also there when you are asleep, which matters in crypto, where Obsidiate’s 50 listed coins trade around the clock and the worst candles have a documented fondness for 3 a.m.

The honest case for mental stops is narrow: experienced traders, at screens, in instruments where a momentary wick through a level is common and the spread is wide. Even then, most of them keep a hard stop further out as a catastrophe bound. Flexibility inside a hard limit is a strategy. Flexibility instead of a hard limit is a hope.

Key takeaways

• A stop marks where your trade thesis is invalidated — it is a decision, not a prediction.
• Structure-based stops tie the exit to evidence; percentage stops enforce discipline — calibrate either to the instrument’s normal range.
• Position size derives from stop distance, never the other way around.
• Stops cluster at obvious levels, and clustered stops attract price; place yours beyond the obvious spot and expect wicks.
• A stop guarantees an exit attempt, not an exit price — fast markets can fill worse than the trigger.
• Mental stops fail exactly when they are needed; use real orders, especially in 24/7 markets.