Stop Loss and Take Profit on Hyperliquid: Setup, Pitfalls, and Automation

Traders spend most of their attention on entries, but the distribution of your PnL is written almost entirely by your exits. An entry decides whether a trade starts well; the stop and the target decide what every trade is allowed to cost and what it is allowed to pay. Get the exits systematic and a mediocre entry engine survives; get them wrong and no entry engine saves you.

On a 24/7 leveraged venue this is not philosophy. Hyperliquid does not close overnight, and volatility does not wait for your timezone. A position without a resting stop has one anyway — it is called your liquidation price, and as covered in our liquidation guide, it is the most expensive exit on the menu: forced, at the worst prices, and in the backstop case your remaining margin stays with the liquidator vault.

So the practical question is not whether to use TP/SL orders on Hyperliquid, but how they actually behave — because the mechanics have real edges, and the edges are where money is lost.

Take-profit and stop-loss orders on Hyperliquid are trigger orders: they rest off the book, and when the trigger price is touched they fire an order onto it. Two mechanics define everything downstream.

First, the trigger reference: TP/SL trigger on mark price, not on Hyperliquid's last trade. Mark is a median that blends the CEX oracle, Hyperliquid's own book, and external perp mids — the same manipulation-resistant reference used for liquidations. A wick on Hyperliquid's book alone will generally not fire your stop; a market-wide move will, even if the local book never printed there.

Second, what fires when triggered — your choice of two flavors:

• Market TP/SL converts to a market order with a 10% slippage tolerance. It will almost always get you out, at whatever the book offers within that band.

• Limit TP/SL fires a limit order at a price you set. The gap between trigger and limit is your explicit slippage budget: trigger $10, limit $8 fills like a market order in all but the worst gaps; trigger $10, limit $9.90 may simply never fill if price is moving fast.

That is the fundamental trade-off: market TP/SL guarantees the exit and floats the price; limit TP/SL pins the price and floats the exit. For a stop whose job is making sure you survive, guaranteed-exit is usually the right default — the slippage cost of a market stop is almost always smaller than the cost of a stop that didn't fill.

Hyperliquid gives you two homes for a TP/SL, and they behave differently in ways that matter.

Position TP/SL is set on an open position. By default it covers the entire position — whatever size the position has when the trigger hits, that is what it closes, and it is cancelled automatically when the position is gone. This is the right tool for 'protect this whole trade.' One subtlety: if you set a custom size instead of the default, the order becomes fixed-size and stops tracking the position — scale in afterwards and the untracked remainder is unprotected.

Bracket TP/SL is attached to an entry order at placement (the one-click way to enter with exits predefined). The children are sized to the parent order and go live when the parent fully fills. Cancel a partially-filled parent and the children cancel with it — worth knowing if you work partial fills manually. Together the legs behave as an OCO pair: one exit closes the position and the sibling is cancelled rather than left armed.

Both flavors close exposure rather than flip it — a triggered stop cannot leave you short when you were long. The practical rule: brackets when you enter (so exits exist from the first second), position TP/SL to adjust as the trade evolves, and if you scale in, re-check that your protection still covers the full size. Mismatched stop-size-after-scaling is one of the quietest ways to lose more than you planned.

Every expensive TP/SL story falls into a small set of patterns:

• The gap-through. Price jumps over your stop level into an air pocket. A market stop fills — up to 10% worse than the trigger in the extreme. That tolerance is your worst case; size positions so that even that fill is survivable, not ruinous.

• The limit stop that never filled. The inverse failure, and the worse one: a tight limit stop in a fast market doesn't fill, and you are still in the position, now deeper underwater, with your 'protection' resting above the price. If you use limit stops, the trigger-to-limit gap must be generous enough to fill in exactly the conditions that fire stops — chaos.

• The stop behind the liquidation price. At high leverage there may be less room than your stop needs: if your liquidation sits closer than your stop level, the exchange liquidates you first and the stop never gets its turn. The stop must fit inside the margin geometry — see the liquidation-distance table in our margin guide.

• Mark-price surprises. Your stop can fire without Hyperliquid's book printing the level (a CEX-led move drags the mark), and a local wick through your level may not fire it. Both are the mechanism working as designed — but if you didn't know, both feel like betrayal.

• The manual override. The most expensive failure mode runs on no code at all: watching price approach the stop and cancelling it, 'just this once.' Every liquidation-adjacent blowup contains this step. A stop you can talk yourself out of is a suggestion, not a stop.

Mechanics aside, most TP/SL damage comes from where the levels go and how they get moved:

• Stops go at invalidation, not at pain tolerance. The stop belongs where the trade idea is objectively wrong — below the support or swing low that defined the setup — not at the loss that feels acceptable. If invalidation is further than you can afford, the answer is a smaller position, never a closer stop.

• Targets go at structure. Take profit into the levels where opposing supply actually sits — prior highs, range boundaries, high-volume zones — not at a round number of dollars you would like to make. A target beyond the next real wall is a plan to watch profit evaporate.

• Demand a minimum risk-reward before entry. If the distance to a structurally honest target isn't a multiple of the distance to the stop, the trade fails arithmetic before it ever fails on the chart.

• Move stops one direction only. Tightening into profit (or to break-even once the trade has earned it) is management; widening a losing stop is re-underwriting a worse trade at the worst moment. Make break-even moves at predefined thresholds, mechanically — not when a red candle scares you into it.

None of these rules is controversial. The hard part is that they must be applied at entry time, every time, including when you are most convinced — which is exactly when discretion fails.

The common thread in every failure mode above is sequencing: exits decided after the position exists get decided by a person with a position — biased, hopeful, and negotiating with themselves. The fix is structural: make exits part of the trade's definition, placed the moment it fills, by something that doesn't renegotiate.

That is how Hyperhelm executes. Every governed verdict ships as a complete bracket — entry limit with take-profit and stop-loss as native reduce-only triggers, OCO'd, live from the fill. The levels are not vibes: stops sit beyond the structural support that defines invalidation, targets sit in front of the heavy zones the structure map says will fight back, and the risk gate rejects any intent whose stop distance, risk-reward, or liquidation buffer fails the math from this guide. The stop you would have cancelled at 3am was placed by the system at noon — and it isn't asking your opinion.

You keep the override — it is your account and your signature — but overrides are logged, and the public benchmark scores governed versus ungoverned execution side by side. Discipline you can measure beats discipline you promise yourself. The cockpit is free to open; look at how a fully-specified, pre-bracketed intent reads before you take another naked position.