Most traders pick leverage the way they pick a font size. 10× feels normal, 20× feels a bit spicy, and 50× is for days when you are sure. The number on the slider never says what it actually decides: how far price can move against you before the exchange closes your position and keeps your margin. That distance is simple to compute, and once you have computed it a few times, high leverage stops looking like a choice about conviction and starts looking like a choice about noise.
This article walks through the arithmetic for isolated and cross margin on Bybit and Hyperliquid, shows where traders get hurt by it, and ends with a sizing rule you can apply before the next order.
The one formula worth memorizing
On a Bybit USDT perpetual in isolated margin, the liquidation price of a long is:
liq_long = entry × (1 − 1/leverage + MMR)
liq_short = entry × (1 + 1/leverage − MMR)
MMR is the maintenance margin rate. For BTC and ETH at retail sizes, Bybit's first risk tier is 0.5%. Larger positions move into higher tiers with a higher MMR, which pulls liquidation closer.
Strip the entry price out and you get the distance from entry to liquidation as a share of price:
distance ≈ 1/leverage − MMR
That is the whole idea. Leverage sets your margin as a fraction of the position (1/leverage). The exchange liquidates you once losses have eaten that margin down to the maintenance level. Everything else is detail.
What that looks like on a BTC long
Entry at $70,000, isolated margin, MMR 0.5%, fees and funding ignored:
| Leverage | Distance to liquidation | Liquidation price |
|---|---|---|
| 3× | 32.83% | $47,017 |
| 5× | 19.50% | $56,350 |
| 10× | 9.50% | $63,350 |
| 20× | 4.50% | $66,850 |
| 25× | 3.50% | $67,550 |
| 50× | 1.50% | $68,950 |
| 100× | 0.50% | $69,650 |
Read the bottom half of that table slowly. At 50×, a $1,050 move in BTC ends the trade. At 100×, $350 does it. BTC can travel that far inside a single hourly candle, and a wick does not care that the close came back.
Short side is the mirror image: at 20× short from $70,000 the liquidation sits at $73,150.
You can check any row with the Bybit liquidation price calculator, which uses the same formula and lets you change the maintenance margin tier.
Hyperliquid computes it differently
Hyperliquid derives maintenance margin from the asset's maximum leverage: it is half the initial margin at the asset's maximum leverage. So an asset with 40× max leverage has a maintenance rate of 1/(2 × 40) = 1.25%, and one with 10× max has 5%.
The isolated liquidation price for a long becomes:
liq_long = entry × (1 − 1/leverage) / (1 − maintenance_rate)
Two things follow from that.
First, the same leverage gives a different distance on a different asset. At 10× on an asset with 40× max, liquidation sits 8.86% below entry. At 10× on an asset whose max is 10×, it sits 5.26% below.
Second, if you run any asset at its own max leverage, liquidation sits roughly halfway to where 1/leverage would put it. At 40× max and 40× used, the distance is 1.27%, not 2.5%. At 20× on a 20× asset it is 2.56%, not 5%.
The max leverage per asset is on the Hyperliquid order screen. Plug it into the Hyperliquid liquidation price calculator together with your entry, size and leverage before you place the order.
Why "I use low leverage" is often not true
Here is where most of the damage happens. Leverage looks like a risk setting, but on its own it only decides how much margin you post. Your loss on a given price move depends on notional size, and notional is margin × leverage.
Take a $2,000 account and two traders who both "risk" $500 of margin on a BTC long:
- Trader A uses 5×. Notional is $2,500. Liquidation sits 19.5% away.
- Trader B uses 20×. Notional is $10,000. Liquidation sits 4.5% away.
Trader B's liquidation is closer, and every move costs more. A 3% move against them costs $300, or 15% of the account, while the same move costs Trader A $75. Twenty-five percent of the account is gone the moment liquidation hits. The slider was presented as a margin setting, and it quietly quadrupled the position.
Fees and funding follow notional too. Trading fees are charged on the full position value at entry and exit, and funding is charged on notional every settlement window. A 20× position pays four times the fees and funding of a 5× position on the same margin, and on a losing streak that drag compounds. In my own history (940 closed Bybit trades, all-time as of July 2026), fees and funding came to $1,410.69: $841.68 in trading fees and $569.01 in funding. The funding fees article breaks down how that bill builds up.
Cross margin moves liquidation further away and raises the stakes
In cross margin, your whole wallet balance backs every open position. That pushes the liquidation price further out, which feels safer. What changes is what you lose when it hits.
Same $2,000 wallet, $10,000 BTC long in cross, maintenance margin 0.5% of notional ($50), no other positions, fees ignored. Liquidation comes when losses reach roughly $2,000 − $50 = $1,950, a 19.5% move. That is the same distance as Trader A in isolated. The difference is that Trader A loses $500 at liquidation and Trader B in cross loses the account.
Cross is a fine tool for traders who size every position from a stop. It is dangerous for traders who size from the leverage slider, because every oversized position borrows safety from the rest of the account until there is none left.
The check that matters: does liquidation sit beyond your stop?
Before any order, ask one question: will my stop fire before liquidation does?
Work it the other way round:
- Decide where your stop goes based on the chart, not the account. Say 2% below entry.
- Decide how much of the account you will lose if it fires. Say 1% of $2,000, so $20.
- Position size follows: $20 / 2% = $1,000 notional.
- Now pick leverage only to set margin, and check that liquidation is further from entry than your stop, with room for slippage on a fast move.
At $1,000 notional, 5× posts $200 of margin and puts liquidation 19.5% away. 20× posts $50 and puts it 4.5% away. Both are fine: the stop at 2% fires first, and the loss is $20 either way. 50× puts liquidation 1.5% away, inside your stop, and that is a broken trade before it starts.
The position size calculator does steps 1 to 3 and warns you when liquidation lands before the stop.
Once you size this way, leverage becomes a capital efficiency setting. Higher leverage leaves more free balance in the wallet; it does not make the trade bigger. Over-leverage, in practice, means letting the leverage number decide the size.
What I would look at in your own history
If you have a year of perp trades, three questions will tell you more than any rule:
- How many of your losses were liquidations rather than stops? A liquidation means the trade had no exit you chose.
- What was your notional per trade relative to your account? If it swings by 5× from week to week, your risk is being set by mood.
- Did your largest losses come from your highest-leverage trades? If they did, the problem is size, and the fix is the four steps above.
My own all-time numbers include a −$8,825.56 max drawdown (95%) and a worst day of −$3,240.33. The data does not label why each loss happened, so I won't pretend it does.
PerpLog syncs your Bybit or Hyperliquid history read-only and counts fees and funding on every trade, so you can answer those questions from your own fills instead of from memory.
Takeaway
Liquidation sits about 1/leverage away from entry, minus maintenance margin, and on Hyperliquid at max leverage it sits about half that. Leverage itself sets margin, and notional sets the loss. Put the stop where the chart says, size the position from what you are willing to lose at that stop, then choose leverage so liquidation lands well beyond it.
Tools: Bybit liquidation price calculator · Hyperliquid liquidation price calculator · Position size calculator