You are short, the market is falling hard, and your position is deep in profit. Then it is gone. Not stopped out. Not liquidated. Closed by the venue, at a price you did not choose, while you were on the winning side.
That is auto-deleveraging, usually shortened to ADL. It is part of how perpetual futures deal with losses, and it can reach a trader who did nothing wrong.
To see why it exists, you have to follow a loss that a liquidation could not absorb. This piece stays provider-neutral. The exact rules differ from venue to venue, so it explains the logic the common designs share, not one exchange's formula.
Key Takeaways
- A loss that a liquidation cannot cover becomes a deficit, and the winning side of the trade still has to be paid.
- An insurance fund is usually the first backstop: it pays the deficit so that profitable traders are paid in full.
- When the fund cannot cover the gap, auto-deleveraging closes profitable positions on the other side, usually ranked by profit and leverage.
- You can be deleveraged without breaking any rule, so your place in that ranking is worth watching on violent days.
Where the Loss Goes When a Liquidation Fails
Every perpetual futures position has a counterparty. When you are long and the price falls, your loss is a gain for someone on the short side. The venue sits in between and makes sure the winner is actually paid.
Liquidation is the tool for that. Your position carries a maintenance margin, and once losses eat into it, the venue closes the position. Closing early is deliberate. The venue wants the position gone while there is still margin left to pay for the loss.
Beyond the liquidation price lies another price that matters, the bankruptcy price. That is the price at which your margin would be exactly zero. As long as the liquidation fills between the liquidation price and the bankruptcy price, your margin covers the loss, and the winning side is paid out of it.
Now picture the case the system is built to fear. The market gaps through both levels in one move. The book is thin, and the liquidation can only be filled well beyond the bankruptcy price. The margin is gone, the loss keeps going. The difference is a deficit, often called bad debt.
That deficit does not vanish. The winning side still holds a profit that the losing position was supposed to pay. If nobody covers the gap, that profit is unbacked. Every backstop that follows exists to close it.
How much of your account stands behind a losing position depends on your margin mode. Under cross margin, every open position draws on the shared account balance, so more of your money sits between a bad trade and its bankruptcy price. Our article on isolated vs cross margin walks through that trade-off.
The Insurance Fund: The First Backstop
The first line of defense is usually an insurance fund. It is a pool of capital the venue keeps for exactly this situation. When a liquidation leaves a deficit, the fund pays it, and the winning side never notices.
How the fund fills up is a design choice. In a common design, a liquidation that closes better than the bankruptcy price leaves some margin over, and part of that remainder flows into the fund. Over calm periods the fund grows from the liquidations that went well. It is then spent on the ones that went badly.
Designs differ in how they slice it. One approach keeps a single fund for the whole venue. Another keeps a separate fund per market, so a crash in a small market cannot drain the cover for a large one. Other designs add a further layer before the fund: a backstop vault or a group of designated liquidity providers that take over a failing position before it turns into a deficit at all.
The fund is a buffer, not a guarantee. It can only pay out what has flowed into it, and one extreme move can cost more than a long calm stretch has saved. That limit is the reason the next mechanism exists.
When the Fund Runs Dry: How Auto-Deleveraging Works
If the insurance fund cannot cover a deficit, the venue needs another way to keep the market whole. Auto-deleveraging is that way. It is designed as the last resort, after everything else has failed.
The idea is blunt. The bankrupt position still has to be closed, and nobody in the book will take it at a price that covers the loss. So the venue closes it directly against profitable positions on the opposite side. If a bankrupt long has to go, ADL reduces winning shorts. If a bankrupt short has to go, it reduces winning longs.
Those opposing positions are closed, fully or in part, at a price the mechanism sets rather than the book. That price is usually tied to the bankruptcy price of the failed position or to the mark price at that moment. The winners keep their profit up to that price. What they lose is the rest of the trade.
ADL closes winners, not losers. That is the part to keep in mind. The traders who get reduced are the ones the market has been rewarding, because only their open profit can absorb the gap.
Who Gets Picked First: The ADL Queue
ADL does not pick at random. In most designs the venue ranks the opposing positions and starts at the top of that ranking.
The ranking usually combines how much unrealized profit a position carries with how much leverage it uses. A position that is deep in profit on high effective leverage sits near the front. A small position with modest profit and low leverage sits far behind it.
The logic is about risk. A high-leverage winner carries a large exposure relative to its own margin, and it has gained a lot from the very move that broke the other side. Risk comes out where it is largest. Reducing that position first takes exposure out of the system where it is most concentrated.
Your rank moves as the market moves. Taking part of the profit, adding margin or cutting leverage pushes a position further back. Letting a big winner run on high leverage pulls it forward. Where a venue shows an ADL indicator next to your position, it is showing you roughly where you stand in that queue.
The queue matters most on the days that feel best. A sharp move that makes your position a big winner is the same move that can leave someone else's position bankrupt.
What Being Auto-Deleveraged Means for Your Trade
On the receiving end, ADL feels like a forced take-profit. Your position, or part of it, is closed. The profit up to the closing price is realized. The trade you planned is over.
These consequences are worth thinking through before it ever happens to you.
You lose the rest of the move. ADL runs in the most violent part of a market swing, exactly when a well-placed winner may have the most room left. Being closed there can cost more than any fee.
Hedges can break. If your position was one leg of a hedge, ADL removes that leg and leaves the other one standing. A neutral book turns into a directional one, in a market that is moving hard.
You do not choose the price. The closing level comes from the mechanism, not from your order. It is anchored to the failed position or to the mark price, so it is not arbitrary. It is just not the level you would have picked.
You did nothing wrong. ADL is not a penalty and not a rule breach. It is the system protecting its own solvency, and your position happened to sit on the side that could pay for it.
Why ADL Exists Instead of Something Else
ADL is uncomfortable, so it is fair to ask why a venue would use it at all. The short answer: the alternatives spread the same loss in less visible ways.
One older design is the socialized loss. Instead of closing specific positions, the venue takes a share of every winning trader's profit after the fact, often in proportion to its size. Every winner pays something, and nobody knows in advance how much. Profit you thought was yours can be cut after you have already counted it.
Another option is for the venue to absorb deficits from its own balance. That works until a deficit is larger than the venue can carry. It also moves the risk to a place traders cannot see.
ADL makes a different trade-off. It hits fewer traders, but it hits them harder, at a known moment and under a visible rule. A visible rule beats a surprise clawback. You can see your rank, you can change it, and once the mechanism has run, the profit you were paid stays paid.
What to Check in a Venue's Documentation
None of this needs a formula. It needs a handful of answers, and the venue's own documentation is where to find them for the markets you trade.
| Layer | What it does | What to look up |
|---|---|---|
| Liquidation | Closes a losing position before its margin is spent | Which price triggers it, and how much margin is left as a buffer |
| Backstop takeover | Hands a failing position to a designated party before it becomes a deficit | Whether the venue uses one, and who carries the risk |
| Insurance fund | Pays a deficit so the winning side is still covered | How it is filled, and whether it is shared or kept per market |
| Auto-deleveraging | Reduces profitable opposing positions when the fund is not enough | How the queue is ranked, and at which price positions are closed |
The first layer usually runs against the mark price, not the last trade. Why that is so has its own article on mark price vs index price.
If your venue publishes the size of its fund or a record of past ADL events, read that too. It tells you how often the last line of the chain has actually been needed, which no description of the mechanism can.
ADL and a Funded Account
The chain above describes how a venue works. Whether and how a venue-side event reaches a prop firm's account is a question for that firm's own rules, so read those as well.
The rulebook adds context that matters here. Equity is balance plus unrealized P&L, and the equity limits apply to both realized and unrealized losses. Leverage is up to 5x per position, or the maximum Hyperliquid allows for that market if that is lower, and using full leverage is never required.
If a venue-side event like ADL matters to how you trade, settle the bookkeeping question before you start rather than after a volatile day.
Conclusion
ADL is the bill for leverage that the market could not collect from the trader who lost. Liquidation tries first, the insurance fund pays second, and only then are winners asked to give something back. That order is the useful part. It means ADL belongs to thin markets and violent moves, after the other layers have already failed.
It also tells you where the cost lands: on the most leveraged winners of the move. If you carry large open profit on high leverage through a crash, you are standing near the front of the line. Banking part of the gain or cutting leverage is how you step back.
FAQ
Can a stop-loss protect me from ADL?
No. A stop-loss protects against losses, and ADL picks positions that are in profit. If anything, a take-profit order is the closer tool, but it only helps if it fills before the ADL event. Once the mechanism runs, it closes your position at its own price, whatever orders you have resting.
Does isolated margin keep me out of the ADL queue?
Not by itself. Margin mode decides how much of your balance stands behind a losing position. ADL selects winning positions on the other side. Margin mode can still matter indirectly: it changes how much margin backs a position, and that changes the effective leverage that most rankings weigh.
Does ADL only happen in small, illiquid markets?
It is most likely there, because a thin book makes it hard to close a liquidation near its bankruptcy price, and a market with its own fund may have less saved up. A large market can produce a deficit too, if a move is sudden enough that the book cannot absorb the forced closes.
Do I get my position back after ADL?
No. The closed part is settled at the ADL price, and the profit up to that price is yours. To be back in the trade, you open a new position at the current market price, with a new entry level and the usual trading costs.
What happens to the trader whose position went bankrupt?
They lose the margin behind the position. Under cross margin, that can mean the whole account balance. The part of the loss beyond their margin is exactly the deficit that the insurance fund, and if needed ADL, has to cover. The winning side is paid either way.
