Cross Margin vs Isolated Margin: Which to Use

Cross margin and isolated margin are the two ways a derivatives exchange can back a leveraged position with collateral. Isolated margin assigns a fixed amount of collateral to one position, so that position can only lose what you put behind it. Cross margin lets every open position draw on your whole account balance, which delays liquidation but puts the entire balance at risk. The choice shapes how and when you can be liquidated, and it is one of the first settings a trader meets on any perpetual or futures venue covered in our guide to crypto derivatives.

This guide defines both modes, works through an example, and shows how to size a position so your stop-loss is hit long before liquidation.

Quick answer: Cross margin vs isolated margin is the choice between sharing your whole account balance as collateral for all leveraged positions (cross) and ringfencing a fixed amount of collateral for one position (isolated). Isolated margin caps the loss on a position at its assigned margin. Cross margin delays liquidation but can lose the entire account balance.

---

What is isolated margin?

Isolated margin is a margin mode in which a fixed amount of collateral is assigned to a single position, and only that collateral can be lost if the trade goes wrong. The rest of your account balance is untouched by that position.

If you open a long with 1,000 USD of isolated margin, that 1,000 USD is the position's entire budget. When losses eat through it (in practice slightly before, once remaining margin falls to the maintenance requirement), the position is liquidated. Your other funds are not used unless you add margin manually. The main appeal is that the worst case is visible before you enter.

What is cross margin?

Cross margin is a margin mode in which all open positions share the full available balance of the account as collateral. A losing position can keep drawing on free balance, including unrealised profits from other positions, before it reaches liquidation.

Because the pool is larger, each liquidation price sits further away than under isolated margin. The trade-off is that one position going badly wrong can consume the balance meant for everything else, and if the pool runs out, the exchange can liquidate positions across the account.

For a full explanation of forced closures, see what liquidation is in trading.

How does liquidation differ between cross and isolated margin?

Liquidation under isolated margin is triggered when the losses on one position use up that position's own margin, while liquidation under cross margin is triggered only when losses across the account use up the shared balance. Same position, same price move, very different outcomes.

Two terms matter here:

• Initial margin is the collateral needed to open a position. At 10x leverage it is 10% of notional value.
• Maintenance margin is the minimum collateral needed to keep the position open. Below it, liquidation starts. Rates vary by exchange and instrument.

Under isolated margin, the liquidation price depends only on the position. Under cross margin, it also depends on your free balance and every other open position.

Worked example: a 10,000 USD long

Say your account holds 5,000 USD and you open a 10,000 USD long on a crypto perpetual. Fees, funding and maintenance margin are left out to keep the arithmetic clean.

Isolated margin, 1,000 USD assigned (10x):

1. Position notional: 10,000 USD. Margin: 1,000 USD. Leverage: 10,000 / 1,000 = 10x.
2. Every 1% fall in price loses 10,000 x 1% = 100 USD.
3. A 10% fall loses 10,000 x 10% = 1,000 USD, which equals the full margin.
4. The position is liquidated at about a 10% fall. Loss: 1,000 USD. The other 4,000 USD in the account is untouched.

Cross margin, whole 5,000 USD balance as collateral:

1. Same position, same 100 USD loss per 1% fall.
2. At a 10% fall the loss is 1,000 USD, but 4,000 USD of balance is still backing the trade, so there is no liquidation.
3. At a 20% fall the loss is 2,000 USD. At a 30% fall it is 3,000 USD.
4. The position survives until the loss approaches the full 5,000 USD, close to a 50% fall (earlier in practice, once maintenance margin is counted).

The chart shows the two paths. Up to a 10% fall the lines are identical, because the position and the loss per 1% are the same. At 10%, the isolated position is liquidated and its loss stops at 1,000 USD. The cross position keeps going, and by a 30% fall it has lost 3,000 USD, three times the isolated cap.

Neither outcome is better on its own. Isolated margin took a 1,000 USD loss in exchange for a hard ceiling. Cross margin gave the trade room to recover, but if price never comes back, the cross trader loses far more.

Cross margin vs isolated margin at a glance

Why do exchanges use mark price for liquidation?

Exchanges usually trigger liquidations from the mark price rather than the last traded price, so that a single thin-book wick does not wipe out positions. The mark price is typically built from an index of spot exchange prices plus a basis or premium component, and the formula varies by venue (see mark price).

In the chart, the last traded price wicks down to 96.90, below a liquidation level at 98.00, but the smoothed mark price stays above 98.00, so the position is not liquidated. This applies in both modes: the margin mode decides how much collateral backs the position, and the mark price decides when that collateral counts as exhausted. Many venues also let you choose whether a stop-loss triggers on last price or mark price, so know which one you are using.

When does cross margin make sense?

Cross margin makes sense when positions genuinely offset each other, so that a loss on one leg is matched by a gain on another. Examples include:

• Hedged books. A long position paired with a short perpetual future on the same asset, where gains on one side support the margin of the other.
• Several positions managed as one portfolio. A balanced set of longs and shorts can share one pool instead of many small ringfenced ones.
• Traders who actively monitor total account risk, not just each trade.

Cross margin is a poor choice for one large, unhedged trade left unattended. The extra room can turn a small, planned loss into a large, unplanned one.

When does isolated margin make sense?

Isolated margin makes sense for single, high-risk trades where you want the worst-case loss fixed before you enter. Examples include:

• Speculative trades on volatile mid-cap or small-cap coins.
• Event-driven trades around news, where gaps and wicks are likely.
• Learning, while you find out how leverage behaves.

The weakness is a closer liquidation price, so a normal swing can close the trade. The fix is a smaller position, not topping up margin.

Common mistakes with cross and isolated margin

Adding margin to a losing isolated position

Adding margin to a losing position moves the liquidation price further away, which feels like rescuing the trade. In practice it turns a capped loss into an uncapped one, one top-up at a time. Decide the maximum loss before entry and honour it.

Running many correlated positions on cross

Several longs on different coins look diversified, but crypto prices often move together in a sell-off. On cross margin, correlated positions lose at the same time from the same pool, so liquidation arrives faster than any single position suggests. Treat them as one large position when sizing.

Using leverage as the sizing tool

Leverage tells you how much collateral a position needs, not how much to risk. See prop firm leverage for more.

How do you size a position so the stop is hit before liquidation?

You size a position so the stop is hit before liquidation by choosing the stop first, sizing from the amount you are willing to lose at that stop, and then checking that the liquidation price sits well beyond the stop. This works in either margin mode.

1. Set the risk budget. Decide the most you will lose on the trade. Many traders use around 1% of the account. On a 5,000 USD account, 1% is 50 USD.
2. Place the stop from the chart. Put it where the trade idea is proven wrong, not where the loss feels comfortable. Say the stop is 2% below entry.
3. Calculate position size. Position size = risk budget / stop distance = 50 / 0.02 = 2,500 USD notional.
4. Choose margin and leverage. On isolated margin with 500 USD assigned, leverage is 2,500 / 500 = 5x. Ignoring maintenance margin, liquidation sits about 20% below entry (500 / 2,500).
5. Check the gap. The stop at 2% is about ten times closer than liquidation at about 20%. The stop will trigger long before the exchange forces a close.
6. Do not move the stop or add margin once the trade is live.

Compare the worked example above: with the same 2% stop, the 10,000 USD position would lose 200 USD at the stop, four times the 50 USD budget. The problem was position size, not margin mode. See stop-loss placement and risk management in trading for more.

Leveraged crypto derivatives are high risk. Broker disclosures and academic studies consistently find that most retail day traders lose money, and leverage magnifies both the speed and the size of those losses.

How does this apply on a prop trading account?

Velotrade, a multi-asset prop trading firm, runs simulated evaluation accounts with their own risk rules rather than exchange margin modes. On Velotrade, the official hard limits are the daily loss limit and the static maximum drawdown, and position size is limited only by the leverage available for each instrument (for example BTC at 10x during the challenge and 5x once funded).

The daily loss limit resets at 00:30 UTC and is set from the higher of balance or equity: 5% on CLASSIC 2-Step, 4% on CLASSIC 1-Step and 3% on PRO 1-Step. The maximum drawdown is a fixed dollar floor set on the starting balance. Both limits apply to the account as a whole, so the cross margin lesson carries over: correlated positions losing together all count against the same limit. Sizing from the stop is what keeps one bad day from ending an evaluation (see never get liquidated again).

Can AI help with margin and position sizing?

AI tools can help with parts of the process, such as flagging correlated positions, analysing a trade journal for patterns like repeated margin top-ups, or helping backtest trading strategies to see how often a stop would have been hit before a given liquidation distance.

AI does not remove risk. A model fitted to past data can fail when volatility changes. Treat AI trading strategies as an input to human judgement, not a replacement for a written risk plan.