You pull up a liquidation heatmap and the chart lights up. A bright band sits a little above the current price, another one further down. It looks as if the market has shown you its cards: this is where the leveraged positions are, and this is where they will be forced out.
That reading gets the shape right and the certainty wrong. The bands do mark price levels where, under certain assumptions, many positions could be closed by force. But nobody handed the map maker those positions. They were inferred. The map is an estimate, and the assumptions behind it decide what you see.
This piece stays provider-neutral. A heatmap tool can pick its own inputs and its own math, so the article explains the shared logic, not one tool's recipe.
Key Takeaways
- A liquidation heatmap estimates where forced closes could bunch together, built from open interest, assumed entry prices and assumed leverage.
- The map does not show real positions or real liquidation prices; it shows the result of the provider's assumptions.
- Bands above the current price usually stand for shorts at risk, bands below the price for longs.
- A band can matter because forced closes are usually sent as market orders, which can push price toward the next band.
- Your own risk is set by your position and your account limits, not by a band on someone else's map.
What a Liquidation Heatmap Shows
A liquidation heatmap is a chart with price on one axis and time on the other. Over the candles sits a layer of color. Brighter, warmer areas mark price levels where the model expects a larger amount of positions to reach their liquidation price. Darker areas mark levels where it expects little.
Read it like a weather map for leverage. It tells you where the model thinks pressure has built up. It does not tell you that a storm will arrive, or when.
The idea comes from how perpetual futures work. A leveraged position has a price at which its margin no longer covers the maintenance requirement, and at that price the venue closes it. When positions are opened near the same prices and with similar leverage, their liquidation prices bunch together. Those bunches are what the map tries to draw.
The result is attractive because it turns positions you cannot see into a picture you can read at a glance. That is also its weakness. A clean picture hides a rough calculation. To read the picture well, you need to know how the calculation works.
How the Estimate Is Built
No venue hands out a ready-made map of every trader's liquidation price. A heatmap provider has to reconstruct it, and that reconstruction usually rests on three inputs.
The first is open interest, the total size of positions still open on a market. When open interest rises, new positions have been opened. The model takes that rise as a sign that fresh leveraged exposure now exists somewhere.
The second is an assumed entry price. The model usually cannot see where each new position was opened, so it places the new open interest at the prices where trading happened while open interest grew. That is a guess. A reasonable one, but a guess.
The third is assumed leverage. From an entry price and a leverage level, the model calculates where a slice of positions would be liquidated. Since it cannot see the real leverage either, it usually spreads the open interest across a set of typical leverage levels and draws a band for each. Guess on guess.
That makes every band a stack of assumptions. Change the leverage set, the entry logic or the time window, and the same market produces a different picture. Two tools showing the same market can disagree for exactly this reason, and neither of them has to be broken.
Why the Real Liquidation Price Is Harder to Know
Even for a position you can see on your own screen, the liquidation price is less settled than it looks. Hyperliquid is the venue CarrotFunding executes trades on (our Hyperliquid venue page has the details), and its documentation shows why a reconstructed map can only be a rough sketch.
On Hyperliquid, liquidations use the mark price, which combines external CEX prices with Hyperliquid's own book state. The trigger is therefore not simply the last trade on the chart, and a map built from traded prices starts from a different number than the venue does. Our piece on mark price vs index price explains why venues lean on that reference.
The same documentation notes that the actual liquidation price of a cross margin position does not depend on the leverage the trader set: a cross position at lower leverage simply uses more collateral.
For an isolated margin position, the leverage setting does matter, because it decides how much margin is allocated to the position. A model that assigns a single leverage level to a slice of open interest cannot tell these cases apart.
And the level moves after the position is open. Hyperliquid's documentation says the liquidation price shown for an open position may still not be the actual one, because funding payments or changes in unrealized P&L on other cross positions can shift it. A model that fixes each band at the moment of entry follows none of that.
A map built on the open interest figure cannot remove this. The inputs the model would need, every position's margin, mode and account, are not part of that figure. A heatmap can only be as precise as the data it is fed.
How to Read the Map
Once you treat the map as an estimate, reading it gets easier. Read the assumptions before the colors. Four things deserve a look before any bright band does.
- Which side of price a band sits on. A long is liquidated when price falls, so estimated long liquidations sit below the current price. A short is liquidated when price rises, so estimated short liquidations sit above it.
- The time window. If your tool lets you choose how far back the model collects open interest, that choice shapes the map. A short window shows recent positioning; a long one can include positions that were closed long ago.
- The color scale. Intensity is relative to what else is on the screen. The brightest band on a quiet day can stand for far less exposure than a middling band on a busy one, so check the scale, not just the color.
- What price has already crossed. When price runs through a band, the positions the model placed there would, by its own logic, have been liquidated. Whether the map clears that band afterwards depends on the tool, which is one more reason to know how yours works.
None of these checks needs special access. They only need you to remember that every color on the screen was calculated, not observed.
What a Heatmap Is Not
A heatmap gets mistaken for three things it is not.
It is not raw exchange data. The bands come out of a model. A venue records actual positions and actual liquidations, but the map is a calculation made by whoever draws it, with their choices built in.
It is not a forecast. A band says that forced closes could be concentrated at a level if price gets there. It says nothing about whether price will get there. Price can drift away from a bright band and never come back.
It is not a trading signal. One reading treats bright bands as magnets that price has to visit. The map cannot carry that weight. It shows where the model places risk, not where the market is headed, and no part of the calculation looks at what traders will do next.
None of this makes the map worthless. It is context, not a call. It gives you one view of how leverage may be distributed, with its method baked into every pixel.
Why Clusters Can Still Move Price
There is a real mechanism behind the interest in these maps. A liquidation is not a quiet bookkeeping entry. On Hyperliquid, when an account's equity falls below the maintenance margin, the venue first tries to close the positions by sending market orders to the book. Market orders take liquidity, and taking liquidity can push price further in the same direction.
If enough positions sit close together, that push can reach the next group of liquidation prices, whose market orders push further again. Then it feeds itself. That chain is what a liquidation cascade means, and it is the reason a map of possible bunches gets attention at all.
A cluster is fuel, not a spark. Something else has to move price into the band first. The map can suggest where fuel may lie. It cannot tell you what lights it, and it cannot tell you whether its estimate of the fuel is right.
Checking the Estimate on an On-Chain Venue
On a venue that keeps its order book in a private system, positions and liquidations stay inside that system, and an outside model has little to test its guesses against. On a venue whose order book runs on-chain, orders, fills and liquidations become part of a public record.
That changes the question you can ask. Instead of taking the colored bands on faith, you can look at what actually happened when price crossed a level: did liquidations occur there, and in what size? The record gives you a way to test whether a map has been describing the market well.
This is the trust-to-verify habit applied to a chart. A heatmap is a claim about the market. When the underlying record is public, you do not have to accept that claim as given; you can hold it against what the chain shows. Verify, then decide.
Using a Heatmap Without Letting It Trade for You
If you use a heatmap at all, use it to ask better questions about your own position, not to pick a direction.
Start with your own liquidation price, the one your venue shows for your actual position, and remember that it can move with funding and with the rest of your account. That level is yours. A band on a map is about an estimate of everyone else.
Then look at where your stop sits. If it rests inside a level where the model expects heavy forced selling, your stop can trigger in the same sweep.
A stop that becomes a market order fills against whatever liquidity is left, which can be worse than your stop price. That gap is slippage, and in a cascade the liquidity left in the book can be thin.
On a challenge, a different line matters as well. On CarrotFunding, the Maximum Daily Loss limit is the lowest equity level an account can reach in a single trading day before breaching, and equity is balance plus unrealized P&L. A breach is final, as our piece on what happens when you breach explains, so a fast sweep through a band can put you on that line before you have time to react.
Conclusion
A liquidation heatmap is worth exactly as much as the assumptions behind it. Treat it as a reconstruction: open interest placed at guessed entries and spread across guessed leverage, drawn as color. That makes it good for one narrow job, seeing where forced closes could pile up if price travels there, and of no use for predicting where price goes next. The real decision in front of you is not whether to trust the map. It is whether you know your own levels, your liquidation price, your stop and your account limit, well enough that the map only adds context. Where the venue's record is public, you can go one step further and check the map against what happened.
FAQ
Is a liquidation heatmap the same as an order book heatmap?
No. An order book heatmap draws resting limit orders that traders have actually placed, as they appear in the book over time. A liquidation heatmap draws estimated liquidation prices of open positions, calculated by a model. The first shows orders that exist. The second shows where forced orders could appear if price gets there.
Does a heatmap cover positions on every exchange?
Only the positions behind the data the provider feeds in. A map built from one venue's open interest knows nothing about positions elsewhere, and a map that combines venues has to merge figures that each venue reports in its own way. Before you read a band as the whole market, check which venues the tool covers.
Does a liquidation heatmap include stop-loss orders?
Not as such. A stop is an order a trader chooses to place; a liquidation is a forced close that follows from margin. A model built on open interest and assumed leverage has no view of where stops rest. Stops near a band can still add to the same move once they trigger, which is one more thing the map cannot show you.
Can traders set off a cascade on purpose?
Pushing price into a band takes real market orders, and those orders pay the spread and slippage on the way. Whether liquidations then follow depends on positions the pusher can only estimate. Venues that check margin against a mark price rather than the last trade also make a single sharp print less useful as a trigger.
