Why do exchanges use "mark price" instead of the last traded price to calculate liquidation thresholds?
A single exchange can occasionally show an extremely brief price spike — often caused by a large market order momentarily punching through a thin order book, producing an anomalous quote the industry calls a "scam wick." If liquidation thresholds were tied directly to that instant's last traded price, a single wick could wrongly force-close an otherwise safe position.
Mark price is a volume-weighted average across multiple major spot exchanges, used to filter out anomalous moves on any single venue so liquidation decisions track the market's true level rather than one exchange's momentary noise. But this mechanism solves the problem of "will I be wrongly triggered" — it does nothing to solve "once I am genuinely triggered, at what price will my position actually be sold." That price still depends on the real order book depth at that moment, which mark price offers no protection against.
Why does a liquidation cascade self-reinforce, and is there a way to gauge the risk in advance?
A liquidation cascade self-reinforces because forced liquidation is itself a market sell order (for liquidated longs) or a market buy order (for liquidated shorts), which directly moves the price — and that price movement can then trigger the next batch of leveraged positions to hit their own liquidation threshold, creating a chain reaction. How long and how violent that chain becomes depends on how densely liquidation levels are "stacked" within a given price range: if the market has been one-sidedly bullish for a while, leveraged long positions tend to cluster around similar liquidation prices, and once the price starts falling, that thick, dense cluster produces an especially violent cascade.
For an average trader, precisely locating those liquidation clusters requires specialized on-chain or exchange positioning data tools, which is a fairly high bar. A simpler warning sign, though: if the market has been rising one-sidedly for a while and funding rates have stayed persistently elevated, that signals long-side crowding may already be high — and a decline in that environment is more likely than usual to turn into a liquidation cascade.
Can large orders actually avoid slippage by splitting into smaller pieces? How do institutions handle this?
Splitting orders genuinely reduces slippage per execution, since each smaller order only "eats" a shallower layer of the order book rather than sweeping through the entire depth at once. But it's not a complete solution — if the split orders fire too fast and too close together, the market can still feel sustained one-directional pressure, and market makers may start pulling quotes or repricing in response, diminishing the slippage improvement. If the intervals are too long, you take on the risk that the price keeps moving against you while you wait.
Institutional-grade approaches typically use algorithmic trading to dynamically adjust the pacing of order splits, while also referencing order book depth across multiple exchanges to spread execution rather than concentrating all volume on a single venue. That infrastructure is a fairly high bar for retail traders, but the core logic — don't sweep your entire position through with a single market order — is a principle retail traders can borrow directly.
If I'm not a high-frequency trader and only enter or exit occasionally, how do these ideas actually apply to me?
Most non-high-frequency traders don't need to learn algorithmic order-splitting, but three practical habits are worth keeping. First: for non-urgent positions — say, you were already planning to scale in or out gradually — prefer limit orders, which save on fees and avoid slippage risk entirely. Second: if you genuinely must use a market order (say, you need to exit immediately), glance at order book depth or recent volatility first — the thinner and more violent the market, the larger the gap between your expected execution price and what's shown on screen is likely to be, so prepare for that mentally. Third: if you hold a leveraged position, set your liquidation price with real buffer, since you can't control whether you'll end up sitting inside the next dense liquidation cluster — and once forced liquidation happens, your position is sold at market price, with no room to negotiate.
Most traders comparing exchanges start with the fee schedule — what's the maker rate, what's the taker rate, is there a native-token discount. That instinct isn't wrong; fees are a quantifiable, optimizable cost over the long run. But that instinct misses something: fees are a known, fixed cost, while slippage is an unknown cost that expands precisely when things go worst. These are not the same order of risk.
Over 48 hours from June 4 to June 6, 2026, Bitcoin fell from roughly $67,000 to $59,100, and more than $3 billion in leveraged derivatives positions were forcibly closed — in the single worst session, long positions accounted for nearly 85% of all liquidations. This wasn't a random crash; it was a textbook liquidation cascade: falling prices triggered a batch of leveraged long positions to be forcibly closed, the exchange's risk engine sold those positions at market price, that additional selling pressure pushed prices lower still, which triggered the next batch of longs to hit their liquidation threshold — repeating until the densely packed cluster of liquidation levels in that price range was cleared out.
Traders being forcibly liquidated never had the option of placing a limit order — the system closes the position with a market order directly, meaning that during the steepest minutes of the cascade, sell orders flooding into a thinned-out order book execute at prices considerably worse than the "current price" shown on screen. That gap is slippage, and it happens at exactly the moment you have zero negotiating power.
Back to everyday trading: for most exchanges' spot markets, the base rate under the maker-taker fee model typically sits around 0.08%–0.10%. Even without any optimization at all, this cost is fixed and calculable in advance — you know exactly what you'll pay before you place the order. Slippage is entirely different: when liquidity is ample and the market is calm, slippage on a market order can be negligible; but during sharp volatility, when order book depth thins out in an instant, the actual average execution price of the same market order can end up several times — even more than ten times — more expensive than the fee itself.
This is also why derivatives exchanges have widely shifted to using "mark price" — a volume-weighted average across multiple spot exchanges — rather than a single exchange's last traded price to calculate liquidation thresholds. A momentary price spike on one exchange (a so-called "scam wick") shouldn't trigger your liquidation directly. But this mechanism protects against being wrongly triggered into liquidation — it does nothing to protect the actual execution price once liquidation does happen. Slippage risk is still very much alive at that moment.
At its core, slippage happens because your market order "eats" through the order book layer by layer — up or down — until it accumulates the volume you need. The thicker and denser the order book, the smaller the slippage; the thinner the book, the larger the slippage. The problem is that order book depth isn't fixed: a normally liquid trading pair can see market makers (the people posting those maker orders) pull their quotes and step back the moment panic selling begins, thinning the book right at the exact moment you most need to exit and are most reliant on a market order. This is the most insidious part of slippage — it isn't a fixed tax rate, it's a tax rate that gets amplified precisely when you're most vulnerable.
If you only optimize for fees while ignoring slippage, you're carefully saving a known, small amount of money while leaving yourself completely exposed to an unknown, potentially much larger one. Practical adjustments include: for non-urgent position entries and exits, prefer limit orders, which both capture the lower maker rate and give you control over your execution price; if you must use leverage, set your liquidation price with meaningful buffer rather than trading right up against the edge, since you have no way of knowing whether you'll end up sitting inside the next dense liquidation cluster; and for large positions, consider splitting orders into smaller batches rather than sweeping the entire order book with one market order. Fees determine how efficient your trading is over the long run — slippage determines whether your account survives those few extreme minutes.