How to Avoid Liquidation in Crypto Futures: the Leverage and Margin Math
Why liquidation price is a function of leverage and maintenance margin, an illustrated example at 10x, and why the stop — not the liquidation engine — should be the exit that fires.
Liquidation is arithmetic, not bad luck
A perpetual futures position is liquidated when its margin can no longer cover the exchange's maintenance requirement. Nothing about that event is random: it is a function of your leverage, the venue's maintenance-margin rate, and fees. Traders who get liquidated 'unexpectedly' almost always sized the position without ever computing where that line was.
The approximate shape, for an isolated long: the liquidation price sits roughly at entry × (1 − 1/leverage + maintenance rate). Higher leverage pulls the line toward your entry; the maintenance rate pulls it slightly closer still.
An illustration at 10x
Entry $60,000 on a BTC perpetual, isolated margin, 10x leverage, and a venue maintenance rate of 0.5%: the line falls near $60,000 × (1 − 0.10 + 0.005) = $54,300 — about 9.5% below entry. At 25x the same account of arithmetic puts it around 3.5% below entry; at 50x, under 2%. An ordinary hourly wick covers 2%.
Treat every such number as an illustration, not a quote. Real venues tier the maintenance rate by position size, funding accrues, and mark price — not last price — is what triggers the engine. The exact line comes from your venue's own rules; a serious tool verifies those before quoting a number, and so should you.
The stop should fire first
The practical rule that falls out of the math: your stop-loss must sit between your entry and the liquidation line, with room to spare. If the liquidation line is closer than your intended stop, the position is over-leveraged for that trade — the exchange's engine, with its liquidation fee, has become your effective stop.
Working backwards is safer than working forwards: choose the stop from your analysis, choose the risk budget from your account, and let those two derive the size and the leverage. Leverage chosen first is a conclusion in search of an argument.
- Decide the stop price from structure, not from the margin left.
- Compute size from risk budget ÷ stop distance — the same formula as any instrument.
- Check the implied liquidation line sits well beyond the stop.
- If it does not, reduce leverage or size — the trade idea has not changed, only its financing.
Funding and fees eat the buffer
A position held across funding windows pays or receives the funding rate every interval — on a leveraged position, an apparently small rate compounds into real margin drain. Entry and exit fees subtract from the same buffer. A position that would survive the price path on paper can still be liquidated by the slow leak of costs it was never sized to carry.
Run the numbers yourself
Review a perpetual trade before you open itCommon questions
Does cross margin prevent liquidation?
It postpones it and raises the stakes: cross margin lets the whole wallet defend one position, so the line moves further away but what is lost when it is reached is no longer one position's margin. Isolated margin caps the damage at the position; cross trades that cap away for distance.
Is a lower leverage always safer?
Lower leverage moves the liquidation line further from entry — that part is pure arithmetic. Whether the trade is sound is a different question the line cannot answer. A well-placed stop at 5x and the same stop at 2x plan the same loss if size is adjusted; leverage changes where the forced exit is, not whether your reasoning was right.
Why did I get liquidated above my liquidation price?
Engines trigger on mark price, which can diverge from the last traded price you were watching, and liquidation fees are taken from remaining margin. Both effects move the effective line slightly against you — one more reason the stop, not the engine, should be the exit that fires.
Maximelion is decision support, not financial advice. It does not predict markets, recommend trades, or guarantee outcomes. All figures on this page are arithmetic examples, not recommendations.