This article is for educational purposes only. Leverage.Trading is an independent educational and analytics publisher and not a broker, exchange, or investment advisor. Trading with leverage, margin, futures, or derivatives carries a high risk of rapid or total loss. This content is not financial advice and should not be used as a substitute for independent research or professional advice.
Auto-deleveraging (ADL) is a last-resort risk mechanism used by crypto futures exchanges to close profitable positions when the insurance fund cannot cover losses from a bankrupt account. It does not target the failing trader. It targets the trader on the winning side of the trade that created the gap the exchange could not cover.
ADL is not a consequence of bad trading. It is a consequence of being correct at the wrong moment, while someone else’s position collapses beyond what the exchange can handle through normal liquidation.
Auto-deleveraging closes profitable positions, not losing ones. A position generating strong unrealized gains can be force-closed without warning at the bankruptcy price of a failed trader’s account. The closure executes instantly, at a price the selected trader did not choose, and the notification arrives after the position is already gone.
Liquidation closes losing positions. ADL closes winning ones. Auto-deleveraging is the exchange’s final protection mechanism, activating after partial closure, crypto futures liquidation, and insurance fund absorption have all been exhausted.
Rather than targeting the failing trader, it selects profitable traders on the other side and closes their positions at the bankruptcy price of the failed account.
The mechanism is most common in perpetual futures markets, where positions carry no expiry and insurance fund exposure builds continuously across every open contract. Major exchanges report fewer than 0.1% of liquidations result in ADL events under normal market conditions. ADL replaced the earlier shared-loss model, where fund shortfalls were spread across all profitable traders on the opposite side.
What most traders think: ADL closes the positions of traders who took on too much risk or are losing money. What actually happens: ADL closes profitable positions on the winning side of the trade. The trader being deleveraged did nothing wrong.
When a losing position creates a gap the insurance fund cannot cover, ADL fires. The exchange takes over a position that reaches its liquidation price and attempts to close it in the open market. The insurance fund is a reserve pool built from excess margin recovered during successful liquidations.
On Binance, ADL activates when combined bankrupt positions exceed the Maximum Takeover Capacity. This threshold is based on the fund balance. When outstanding losses exceed what the fund can absorb, the engine fires.
When ADL fires, selected positions close at the bankruptcy price of the failing account, not at the current mark price. A profitable BTC long closing at the failed account’s $48,200 bankruptcy price captures none of the $900 gap to the $49,100 market price. The remaining upside is gone the moment the engine fires.
The exchange attempts to close a leveraged position at or above the bankruptcy price when it hits its liquidation price. If extreme volatility prevents the fill, the insurance fund absorbs the shortfall. ADL activates when bankrupt positions exceed the fund’s Maximum Takeover Capacity on Binance, or when Bybit checks every 8 hours whether cumulative losses exceed a threshold.
The engine selects profitable traders from the top of the priority queue and closes both sides at the bankruptcy price. On Hyperliquid’s October 2025 cascade, ADL began 62 seconds after the first liquidation event. It processed 34,983 closures across 162 assets in under 12 minutes.
The ADL priority indicator shows where a position currently sits in the selection queue, but ADL can only activate when the insurance fund reaches its capacity limit. A 5-bar reading places the position in the top 20% of the queue, first to be selected if ADL fires, but only after the fund threshold is breached. During rapid liquidation cascades that exhaust fund capacity, a 5-bar position can be force-closed within seconds.
Every exchange using ADL maintains a live priority queue on each side of a contract. The ranking formula multiplies unrealized profit percentage by a leverage-related factor: a 50% gain at 20x ranks ahead of a 20% gain at 10x, regardless of dollar size. When ADL fires, the engine selects from the top of that queue, working down until the failing position is covered.
The Hyperliquid October 2025 data confirmed an unexpected pattern: median leverage of ADL-affected accounts was 0.20x, with average unrealized gains of 80.58% at selection. High unrealized profit percentage, not high leverage, is the dominant ADL risk factor.
Both Binance and Bybit display a 5-bar ADL priority indicator next to each open position. Five lit bars places the position in the top 20% of the queue, first to be selected if ADL fires.
| Bars Lit | Queue Percentile | Risk Level |
|---|---|---|
| 5 | Top 20% | Highest: selected first if ADL fires |
| 4 | 20–40% | High |
| 3 | 40–60% | Medium |
| 2 | 60–80% | Low |
| 1 | Bottom 20% | Lowest: selected last |
ADL priority is calculated by multiplying unrealized profit percentage by effective leverage. Higher scores rank closer to the top of the queue and face greater ADL risk when the exchange’s insurance fund comes under stress.
| Leverage | Unrealized Profit | ADL Score | Approximate Queue Position |
|---|---|---|---|
| 5x | 40% | 2.0 | Lower half: low immediate risk |
| 10x | 40% | 4.0 | Mid-range |
| 20x | 40% | 8.0 | Upper range: elevated risk |
| 10x | 80% | 8.0 | Same score as 20x at 40% profit |
| 50x | 20% | 10.0 | Top of queue |
These figures use a simplified illustration of the ADL ranking formula. Actual queue position varies by exchange, margin mode, and real-time market conditions. Check your exchange’s ADL indicator for your live queue status.
Binance commits to zero ADL incidents on BTC, ETH, BNB, BTCUSDC, and ETHUSDC USDT-margined perpetuals when combined open interest stays below 10 billion USDT. The threshold was raised from 4 billion USDT in August 2025. If ADL occurs on guaranteed contracts, Binance compensates affected traders from the insurance fund.
Bybit checks cumulative losses every 8 hours. When net losses for a trading pair reach 30% of the fund’s peak balance in that window, ADL fires. Bybit charges a maker fee to the ADL-selected trader, unlike Binance, which waives the fee.
Binance’s ADL Guarantee covers BTC, ETH, BNB, BTCUSDC, and ETHUSDC USDT-margined perpetuals only, while combined open interest remains below 10 billion USDT. Every other Binance contract carries full ADL exposure with no compensation: all COIN-M futures, all USDT-M altcoin pairs outside the guarantee list, and all USDC-margined contracts not named. Bybit offers no equivalent guarantee on any contract.
ADL queue position is driven by two inputs: unrealized profit percentage and effective leverage. Reducing either lowers the probability of selection when ADL fires.
Reducing effective leverage moves a position down the queue immediately, without requiring full closure. Realizing partial profits resets the profit percentage calculation on the closed portion. Over-leveraging a winning position compounds both inputs at once, pushing toward the top of the queue.
Using isolated margin limits ADL exposure to one trading pair rather than the entire account. On cross margin accounts, ADL closure can extend order cancellation beyond the affected pair. A crypto futures calculator can model how position size and leverage interact to affect queue ranking.
Insurance fund conditions determine whether ADL fires at all. Funding rates are a secondary signal: stretched funding in either direction reflects too many traders positioned on one side. High open interest combined with stretched funding are the conditions that usually come before ADL events.
Auto-deleveraging (ADL) is a last-resort mechanism used by crypto futures exchanges to close profitable positions when the insurance fund cannot cover losses from a bankrupt account. Selected positions close at the bankruptcy price of the failing account, not at the current market price.
A profitable position gets selected for ADL because the exchange needs a trader on the other side to absorb losses the insurance fund cannot cover. The priority queue ranks positions by unrealized PnL percentage multiplied by a leverage factor, placing the most profitable first.
ADL executes at the bankruptcy price of the failing account, not the current mark price. The gap between the bankruptcy price and the current market price at execution represents gains the selected trader does not capture.
The ADL priority indicator on each open position shows current queue standing. Five lit bars places the position in the top 20% of the queue, first to be selected if ADL fires.
ADL is the last step in the exchange’s liquidation safety system. It fires when the insurance fund cannot absorb losses from a bankrupt position, and it targets the most profitable traders first, not the most reckless.
The Hyperliquid October 2025 cascade showed how the risk works at scale: accounts at 0.20x median leverage made up 99.4% of the positions force-closed. Being right, with a large enough unrealized gain, is what creates the exposure. The mechanism is structural, not punitive.
Three conditions together signal higher ADL risk: rapid fund drawdown, high open interest, and four or five lit bars on the ADL indicator. Managing queue position and watching fund health before a cascade begins is when it still matters.
This article is published under Leverage.Trading’s leverage trading & crypto derivatives education , an independent risk-first learning system built to help traders quantify and manage risk before trading.
