Take-profit and stop-loss

Take-profit and stop-loss are conditional orders with one job: close a position when price reaches a level you chose. They are always reduce-only — a protective order that could open a position would be the opposite of protection. They come in two forms, and the difference between them is the single most important thing on this page.

​

Attached to an order, or attached to a position

Attached to an order. You set take-profit and stop-loss at the moment you place an entry. They activate when that order fills, and they are sized to it. If the entry is cancelled before filling, the protective orders go with it. Attached to a position. You set them from the position panel, on something you already hold. They are scoped to the position rather than to a quantity. That scoping has a consequence worth stating directly: There is one exception. If the protective order converts to a limit order rather than a market order, its size is fixed at the moment of triggering. From then on it is an ordinary resting order and no longer tracks the position.

​

Configuration

Mark price is the safer trigger for protective orders — it is smoothed and far harder to push than the last print, which matters because the whole point is to not be stopped out by a single anomalous trade. Last price reacts faster and tracks what you see printing.

​

How the pair cancels

Take-profit and stop-loss are linked one-cancels-the-other. The exact behavior depends on whether the fired leg completed. If the fired leg fills completely, the other is cancelled. Position closed, nothing left armed. If it triggers into a limit order that does not fill, the other leg stays active. This is the case people do not anticipate, and it is the correct behavior: you still hold a position, so you still want protection. If the market then reverses far enough to fire the second leg, the unfilled order from the first is cancelled and the second leg triggers against the position as it stands — possibly a different size than the first attempt.

​

They fire once

A protective order triggers exactly once. If the resulting order does not fully close the position, the protocol does not re-arm it for the remainder. This matters because a triggered market order is immediate-or-cancel: in a thin book it may fill part of your position and cancel the rest, leaving you still holding something with no protection armed. The reason for firing once is deliberate. Automatically re-arming into a book that just failed to absorb the order means submitting into exactly the illiquidity that caused the partial fill — repeatedly, in the worst conditions.

​

Slippage on a triggered market order

A market order produced by a protective trigger uses a fixed 10% slippage tolerance, regardless of what you have configured for manual orders. The reasoning: a stop exists to get you out. A tight tolerance carried over from your normal settings would mean the stop fires, fails to fill, and leaves you in the position it was supposed to exit. A wide bound accepts a worse price in exchange for actually exiting. Everything else about market orders still applies — the market’s price band and the depth-derived bound both still cap the fill.

​

Changing them

Protective orders cannot be edited in place. To change a trigger or limit price, cancel and create a new one. The gap between cancelling and re-creating is a window with no protection. In a fast market, set the new order first where the platform allows it, or accept that the window exists and keep it short.

​

Where to go next