What is ADL (Auto-Deleveraging)

The market plunged on October 10, 2025. Crypto markets were rocked in early October 2025 by a rapid, broad sell-off and a cascade of forced liquidations. News and market trackers reported large, concentrated liquidations and sharp price moves that squeezed leveraged traders and created conditions where exchange insurance funds were stressed. Those conditions are exactly when exchanges’ backup risk systems — including Auto-Deleveraging (ADL)

ADL is an emergency risk-control mechanism used by many centralized derivatives exchanges. When a trader is liquidated but the exchange’s insurance fund (or other backstops) cannot fully absorb the loss, the exchange automatically reduces (closes) profitable positions on the opposite side of the liquidated trade — starting from the most “risky/high-ranked” profitable accounts — to cover the shortfall. In short: instead of the exchange eating the entire loss, the system takes a slice from opposite-side winners.

Key parts:

• Trigger: ADL is triggered only when regular liquidation + insurance fund cannot close the gap.
• Ranking / priority: Exchanges compute an ADL ranking (often a function of leveraged return, leverage used, position size, and realized/unrealized PnL). Accounts with higher leveraged returns / higher leverage are ranked earlier for deleveraging.
• Settlement price: When a profitable position is taken over by ADL, it’s typically closed at the bankruptcy price of the liquidated position (or another defined settlement price). Any difference between that settlement and the market helps replenish the insurance fund.
• Outcome: Profitable traders can be forced closed (partial/full) — i.e., winners pay for losers when the insurance cushion is exhausted.

Why ADL matters (risks & market effects)

• It transfers realized losses to profitable traders — profitable positions can be partially/fully closed at an unfavorable price. That’s counter-intuitive to many traders who think “winners never lose.”
• It can amplify volatility — if ADL closes many profitable positions, market depth and liquidity can further deteriorate, extending the cascade.
• It changes bot and strategy behaviour — bots that assume “insurance fund covers bankruptcies” may face unexpected closures if ADL is on the table.

How trading bots are affected and how they integrate ADL awareness

1) Sources bots must monitor

Trading bots that operate in derivatives markets should monitor:

• Exchange announcements / system status feeds (ADL events, maintenance). Exchanges sometimes post ADL occurrences to their status channels.
• Liquidation and open interest feeds (third-party services like Coinglass provide liquidation heatmaps and exchange-by-exchange liquidations). These feeds show where forced closes are happening in real time.
• Funding rates & insurance fund metrics (if insurance fund drops quickly, ADL likelihood rises). Some exchanges expose insurance-fund size or even public metrics.

2) Bot design patterns to handle ADL risk

• Leverage management: limit leverage for bot positions or dynamically reduce leverage during high-volatility windows. Lower leverage reduces ADL ranking and probability.
• Position sizing & diversification: avoid huge concentrated positions on a single contract — smaller footprint lowers ADL rank and reduces forced closure impact.
• Mode switching: the bot switches to a “defensive” mode when market-wide liquidations spike (reduce size, widen stop thresholds, take profits earlier). Liquidation heatmaps and funding spikes are good triggers.
• Monitoring ADL priority lights / API fields: some exchanges expose an ADL priority indicator per position or an API field that ranks users; bots can read this and proactively reduce exposure if the ADL percentile rises. (Bybit/UI show ADL priority “lights”; exchanges differ in naming.)
• Cancel orders & withdraw liquidity: market-making bots can withdraw outstanding limit orders when ADL risk rises to avoid being unintentionally matched into deleveraging flows.

3) Execution differences when ADL is live

• Expectation of bankruptcy-price settlement: bots must assume that if ADL matches happen, closures may occur at bankruptcy price — which can be materially worse than mid-market — and recalculate risk accordingly.
• Fail-safe routines: auto-reduce leverage, close high-risk positions, or pause new position openings when an exchange signals ADL is active.

Short, actionable recommendations for bot operators

• Immediately after big market moves: reduce leverage to safe levels (2–5x depending on strategy), cut unhedged exposure.
• Subscribe to exchange status & liquidation feeds and build watch rules that pause or de-risk strategies when >X% of open interest on a pair is liquidated in Y minutes.
• Avoid holding ultra-large, ultra-high-leverage “winner” positions that put you high in ADL ranking — paradoxically, the more profitable + leveraged you are, the more likely you’ll be targeted.
• Plan for worst-case settlement (bankruptcy price) in PnL models and capital buffers.