You open a chart and see a wick stab straight through your liquidation price. Your position is still open. Another day, the last trade sits comfortably above your level, and your position is gone anyway. Both feel like a glitch. Neither one is.
A perpetual futures screen carries more than one price at the same time, and the one that decides a liquidation is usually not the one printed by the latest trade. Once you know which number does what, a lot of confusing moments stop being confusing.
This piece stays provider-neutral. The exact recipe for each price differs from venue to venue, so the article explains the logic they share, not one exchange's formula.
Key Takeaways
- The last price is simply where the most recent trade happened on that venue; it decides your fills, not usually your liquidation.
- The index price aggregates spot prices from several external markets, so no single order book can move it alone.
- The mark price is anchored to the index and smoothed, and in most designs it is the reference for unrealized P&L and liquidation checks.
- A mark price makes brief wicks expensive to weaponize, but a real, sustained move still reaches your position.
Three Prices on One Perp Screen
Look closely at most perpetual futures interfaces and you will find three different numbers that all claim to be the price. They answer three different questions.
The last price is where the most recent trade on this venue was matched. It is the number your chart is built from, and it is the price your own market orders actually fill against. It is real, and it is also local: it only reflects what happened in one order book, in one moment.
The index price is a reference built outside that order book. It combines spot prices for the same asset from several external markets into a single number. You cannot trade the index on the perp venue. It exists to tell the venue what the asset is worth elsewhere.
The mark price sits between the two. It is derived from the index, adjusted so that it still reflects where the perpetual itself is trading, and smoothed so that a single trade cannot yank it around. The mark price is the referee. In most designs, it is the number used to value your open position and to decide whether your margin is still enough.
So you have one price for execution, one for fair value, and one for judgment. Mixing them up is where most of the surprise comes from.
Why the Last Trade Is a Poor Trigger
The last price is honest about one thing only: somebody, somewhere on this book, just traded at that level. It says nothing about whether that level means anything.
Order books are uneven. At quiet moments the resting size near the top of the book can be thin, and one large market order can sweep through several price levels in a fraction of a second. The trade prints. The chart draws a long wick. A moment later, price is back where it was.
Now imagine that wick was the trigger for liquidations. Every leveraged position with a liquidation level inside that wick would be closed by force, at the worst possible moment, because of a print that lasted almost no time at all.
It gets worse, because forced closes are themselves market orders. Liquidating a long means selling into the book, which pushes the last price further down and can reach the next cluster of liquidation levels. That chain reaction is one reason open interest matters so much when a level breaks.
There is also a darker version. If liquidations keyed off the last trade, a trader with enough size would have a reason to push a thin book on purpose, trigger the forced selling, and buy the cheap fills on the way back up. That is the wick-manipulation problem the mark price was built to make expensive.
What the Index Price Is Built From
The index price borrows its strength from being somewhere else. Instead of trusting one order book, a venue collects spot prices for the same asset from several external exchanges and combines them.
How they are combined is a design choice. Typical approaches weight the sources, take a median so that one strange quote cannot drag the result, or drop a source that drifts too far from the rest. The details live in each venue's own documentation, and they are worth reading for the markets you actually trade.
The point of the construction is simple. One order book cannot move it. To shift a well-built index, you would have to move the price on several independent spot markets at once, which costs far more than sweeping a single thin perp book.
That robustness has limits, and it helps to know them. An index can only be as good as its sources. If an asset trades meaningfully on only a few spot venues, the index has less to average over. If a source feed lags or stalls, the index lags with it. And the index updates on its own rhythm, which is not the same rhythm as the trades on your screen.
The index price also does a second job. It is the anchor the perpetual is pulled back toward, which is what the funding rate is for. That mechanism has its own article, so it is not repeated here.
How the Mark Price Is Built
If the index is fair value, why not liquidate against the index directly? Because a perpetual does not always trade at the index. In strong markets it can sit at a premium for a while, in weak ones at a discount. That gap is real money for whoever holds the position.
A mark price that ignored the gap entirely would misstate your profit and loss. So most designs start from the index and add a component that reflects where the perpetual itself is trading, then smooth that component over time. Some venues blend several candidate prices and take the middle one. Others use a moving average of the perp's premium over the index. The exact recipe varies, and that is the part you should look up rather than assume.
What the different recipes share is the effect. The mark price follows the market, but it follows it with some patience. A single print far from the index barely moves it. A move that holds, and shows up on the external spot markets too, moves it fully.
That is why the mark price, and not the last trade, is the usual input when a venue checks your margin. The liquidation level you see on a position is, in most designs, a level the mark price has to reach.
What the Mark Price Protects You From, and What It Doesn't
It is tempting to read all of this as a safety net. It is a narrower thing than that, and the difference matters when you size a position.
The mark price protects you from the brief and the local. A thin-book wick on one venue, a fat-finger order, a burst of forced selling that lasts a moment: these tend to show up in the last price far more than in the mark price. Your liquidation check waits for the reference to agree.
It does not protect you from a move that is real. If the asset falls on every major spot market, the index falls, the mark price follows, and your margin is measured against that. Smoothing delays a real move, briefly. It does not cancel it.
There is also a trade-off you accept along with the protection. Because the mark price follows the market with some delay, it can sit away from the last trade during fast moves. You may watch the chart recover while your margin is still being measured against a mark price that has not caught up yet, or the other way round.
And the mark price says nothing about your fills. A stop that triggers into a thin book still fills at whatever the book offers. If your stop is set to trigger on the last price, a wick can still set it off, even though the same wick would not have triggered a liquidation.
Reading Your Screen With Three Prices in Mind
None of this requires a formula. It requires knowing which number to look at for which decision.
| Price | Built from | What it usually decides |
|---|---|---|
| Last price | The most recent trade on this venue | Your fills and what the chart draws |
| Index price | Spot prices from several external markets | The fair-value anchor the perp is pulled toward |
| Mark price | The index, adjusted for the perp's own market and smoothed | Unrealized P&L, margin checks, and liquidation |
A few habits follow from that split.
Measure your distance to liquidation against the mark price, not against the last candle. When the two sit apart, the mark price is the one your margin answers to.
Check what your stop triggers on. If your interface lets you choose between last price and mark price as the trigger, that choice decides whether a short wick can close your position. Neither option is free: last-price triggers react faster, mark-price triggers ignore more noise.
Read the venue's own description of its index and mark price once, for the markets you trade most. It tells you how many sources feed the index, how the mark price is smoothed, and what happens if a source fails. A short read beats a long evening of wondering why a position closed.
Expect the gap between mark and last to widen when the market is fast. That is the moment to be most careful with tight margin, because the chart in front of you and the number your margin is checked against are briefly telling two different stories.
If you trade on a firm's account rather than your own, the same question applies one level up. A prop firm's loss limits are measured on its own terms, and it is worth knowing which price feeds them. On CarrotFunding, equity is balance plus unrealized P&L, and the equity limits apply to both realized and unrealized losses, so the value of your open positions counts toward them.
Conclusion
The useful shift is not memorizing a formula. It is accepting that the price on your chart and the price that judges your margin are two different instruments, built for two different jobs. The tape tells you what just happened on one book. The mark price tells you what the venue believes the position is worth, slowly and on purpose.
That design makes a cheap wick a poor weapon, and it makes your liquidation level a property of the broader market rather than a single trade. The trade-off is lag. If you size positions so that a lagging mark price cannot surprise you, the mechanism works for you instead of against you.
FAQ
Can I be liquidated if the last price never reached my liquidation price?
Yes, in most designs. The check runs against the mark price, and the mark price can move past your level while the last trade on your venue has not. This happens most often when the perp trades away from the external spot markets, so the index pulls the mark price somewhere the local tape has not gone yet.
Why doesn't my unrealized P&L match the last price on the chart?
Because in most designs open positions are valued at the mark price, not the last trade. In calm markets the two sit close together and you barely notice. In fast markets they can drift apart, and your displayed P&L follows the mark price while the chart follows the trades. Neither number is wrong; they measure different things.
Can the index price itself be manipulated?
It is much harder than moving one order book, because an attacker has to move several independent spot markets at the same time. It is not impossible. An index built on only a few thin sources is weaker than one built on many deep ones, which is why the list of sources in a venue's documentation is worth a look.
Does a mark price make liquidations safer for everyone?
It makes them harder to trigger with a brief, local spike. It does not make leverage safer. When the broader market moves and stays moved, the mark price follows, and positions with thin margin are closed regardless. The protection is against noise and manipulation, not against being wrong about direction.
Is the mark price the same as the price used for funding?
No. Funding is driven by the gap between the perpetual's trading price and the index price, and it is paid between longs and shorts. The mark price is used to value positions and check margin. Both lean on the index, which is why they are easy to confuse, but they do different jobs.
