Stop-limit orders: two prices, and the gap they can fall into

A stop-limit order carries two prices. The first, the trigger, is the level that brings the order to life; the second, the limit, is the worst price you will accept once it is alive.

Until the market touches the trigger, the order does nothing and sits nowhere visible. The moment it is touched, the order is converted into a plain limit order at your limit price, and from then on it behaves exactly like any other limit order: it fills at your price or better, or it does not fill at all.

Why the order type exists in the first place

Two older order types each have one failure.

A limit order is precise about price and silent about timing. It rests in the book immediately and fills only at your number or better. Useful, but if you place a sell limit below the current price to protect a position, it fills straight away at the current better price, which is not protection, it is just selling now.

A stop-market order is precise about timing and silent about price. It waits for a level, then sends a market order. It will exit you. It will not tell you where. In a thin book, a market order sweeps through whatever is available and your average price is whatever the book happened to hold. That is slippage, and in a violent move it can be far from the level you had in mind.

The stop-limit is the bolt of those two together. The trigger supplies the timing. The limit supplies the price discipline. You get to say both “not until price reaches here” and “but not worse than here.”

What you give up is the guarantee of execution. That trade is the whole order type. A stop-market guarantees a fill and not a price. A stop-limit guarantees a price and not a fill.

What the two prices actually do to you

Suppose, with made-up numbers, you are long an asset that is trading at 100 units. You want out if it breaks down, and you do not want to be dumped out at any price whatsoever. So you place a sell stop-limit: trigger at 95, limit at 94.

Three things can happen.

Price drifts down to 95. The trigger fires. A sell limit at 94 goes into the book. Bids are sitting at 94.8, 94.5, 94.2, all above your limit, so you fill near the top of that stack. Say the fill averages 94.7. On a made-up position of 50 units of size, you exited 5.3 below your entry per unit, or 265 units of currency of loss on paper. Unpleasant, planned for.

Price drops to 95 on thin volume and the only bid left is 93.9. Your limit says 94 or better. 93.9 is not better. Nothing fills. Your sell limit at 94 now rests in the book, above the market, which makes it a target for a bounce rather than an exit. If price recovers through 94 you get filled on the way back up, which is fine. If it keeps falling you hold the position all the way down with an order parked somewhere you no longer want to sell.

Price does not drift at all. A news print lands and the asset trades 100 then 91 with nothing in between. Your trigger at 95 was crossed, so the order activated, and your limit order at 94 went into a book where the best bid is 91. No fill. You are still long, and you are 9 units worse off than when you placed the order.

The third case is the one people are not holding in their heads when they place the order.

The three exit orders, side by side

Read the last column as the real question. You are not choosing an order type. You are choosing which failure you would rather live with.

The mistake: treating it as a safer stop

The specific error is setting the limit tight against the trigger and then believing the position is covered.

Trigger at 95, limit at 94.95, is a popular shape because the arithmetic looks tidy and the maximum loss looks capped at a neat number. It is not capped. The limit does not cap your loss. It caps the price at which you are willing to realise a loss, which is a different thing entirely. If the market never offers 94.95, the order does not protect you from anything. It just declines to trade while the position keeps running.

The second version of the mistake is forgetting that the order survives the event. After a gap, you still have a live limit order resting at a price the market has left behind. It does not cancel itself. It sits there, and if price retraces to it later you are filled at a moment you were no longer thinking about that order, possibly after you have already decided to hold.

So the useful way to reason about the gap between trigger and limit: it is a tolerance for how far price can travel in the instant between activation and execution. Wide gap, high chance of filling, less control over the price. Narrow gap, tight price, high chance of no fill. Zero gap, meaning limit equal to trigger, is the tightest and the most likely to leave you holding the position.

And there is a case where no fill is the correct outcome. A single bad print, a momentary wick, a brief dislocation in a thin book: price touches your trigger, finds no real liquidity, and snaps back within seconds. A stop-market order would have sold you into that hole at whatever was there. A stop-limit refuses, nothing happens, and when the book repairs itself you still hold the position at a sane price. The order protected you precisely by not filling. That is not a malfunction. It is the design.

Which means the decision is about what you are defending. If you are defending a price, because the asset is illiquid or prone to false wicks, the stop-limit is the right instrument and no fill is an acceptable outcome. If you are defending against being in the market at all, because the position is leveraged and the gap risk is what would damage you, the limit price is the wrong thing to protect and a stop-market is the honest choice.

Sizing is the quieter half of this. A position small enough that a gap through your stop does not matter changes the whole calculation, which is what the position size calculator is for, and if you are running leverage, the liquidation price calculator tells you where the gap stops being survivable. More terms in the glossary.

Education only, not financial advice, and trading carries a risk of loss.