How liquidation cascades turn a small drop into a crash

You open your phone and Bitcoin is down 12% in an hour with no news to explain it. Nine times out of ten, that's not selling. That's forced liquidations feeding on themselves.

What actually triggers a cascade

A liquidation cascade starts with leverage, not sentiment. Traders borrow to open positions bigger than their collateral, and exchanges set a price where that collateral gets seized and the position closes automatically.

A trader deposits $1,000 and opens a 10x long, controlling $10,000 of BTC
If price drops about 9-10%, the exchange auto-closes the position to protect the loan
That forced close is a market sell order, regardless of what the trader wants

One liquidation barely moves price. Thousands of them, clustered near the same level, do not.

Why does one liquidation cause more liquidations

This is the cascade part. Each forced sell pushes price down further, which drags the next batch of leveraged longs to their own liquidation price.

Price drops 3% on normal selling
That 3% triggers the first wave of 10x-15x longs
Their forced selling pushes price down another 4-5%
That wave hits the liquidation price for 5x-8x longs, and so on

The math compounds because leverage concentrates liquidation prices in clusters. Data from Coinglass regularly shows hundreds of millions in liquidations within a single hour when these clusters get hit, on both longs and shorts depending on direction.

What makes it worse than a normal sell-off

Three things turn a dip into a crash instead of a shrug.

Thin order books: crypto exchanges, even large ones, often have less resting liquidity than people assume, so a $50 million sell order can move price 3-4% on its own
Auto-deleveraging and insurance funds: when a liquidation can't fill at a fair price, exchanges socialize the loss or auto-close opposing positions, adding more forced selling
Cross-margin contagion: traders using one collateral pool across multiple positions can see an unrelated asset's drop force liquidation of a completely different position

The August 2024 unwind is a clean example. BTC fell from around $65,000 to under $50,000 in about 72 hours as a yen carry-trade unwind triggered leveraged crypto longs to cascade alongside a broader risk-off move, per data cited by The Block. The initial move was macro. The size of the crash was leverage.

The part traders miss

Liquidation price isn't your entry price plus a cushion. It moves against you the moment funding rates, fees, or partial liquidations eat into your margin, so your real buffer is often smaller than the number on the screen.

Funding rate payments on perpetual futures slowly drain margin in a trending market, even if price hasn't moved much
Exchanges often liquidate in stages, taking partial closes that still cost fees and slippage before the position fully unwinds
Liquidation price shown pre-trade assumes no funding, no fee drag, and calm order books, all three of which fail exactly when you need them

What to check before you're on either side of this

If you're holding leverage, or thinking about buying a dip that might not be done dipping, a few things are worth confirming first.

Check aggregate liquidation heatmaps (Coinglass, Hyblock) to see where large clusters of leveraged positions sit relative to current price
Look at open interest alongside price: rising OI into a rally often means more leverage waiting to unwind, not organic conviction
If you're using cross-margin, know which positions share collateral, a liquidation in one can trigger a liquidation in another you didn't touch
Treat "no news" crashes as a leverage signal first, not a fundamentals signal, then reassess once forced selling exhausts itself

None of this tells you where price goes next. It tells you why the move was as violent as it was, and whether the next 5% drop has the same fuel sitting underneath it.