
How to read a liquidation heatmap: what the clusters show, and what they do not
Updated November 2026
Updated November 2026
A liquidation heatmap is a chart that estimates where leveraged futures positions would be forced to close, drawn as coloured bands across price and time: the brighter the band, the more estimated liquidations at that price. To read one, find the bright bands nearest the current price, note which side they sit on (long liquidations below, short liquidations above), and treat them as model estimates, not as orders waiting in the market or as targets that price has to reach.
This guide walks through an illustrative heatmap level by level, explains how the estimates are built, and lists what the clusters cannot tell you. It assumes you know how perpetual futures work, because most heatmaps are drawn for perpetual contracts.
What a liquidation heatmap shows
The vertical axis is price. The horizontal axis is time. The colour at each point is the estimated amount of liquidations at that price, from dark (little or none) to bright (a lot). One data site's documentation describes its heatmap as the "model-estimated distribution of liquidations on a time–price plane".
A liquidation happens when losses on a leveraged position use up its margin and the position is closed for the trader. The difference between that and a warning is set out in our guide to margin call and liquidation. Because a falling price hurts long positions, estimated long liquidations sit below the current price. Estimated short liquidations sit above it.
A liquidation map is the same idea without the time axis: bars of estimated intensity by price level. The heatmap adds history, so you can see when a band appeared and whether it is still there.
How the estimates are made
No public chart can see every trader's entry price, leverage and margin. A heatmap, whether it is a BTC liquidation heatmap or one for any other contract, is built from what is public, plus assumptions. The usual method has four steps.
- Open interest. Open interest is the number of derivative contracts still outstanding. When it rises while price trades at a level, the model assumes new positions were opened there.
- Assumed leverage tiers. The model spreads that new open interest across leverage settings, such as 10x, 25x, 50x and 100x. One data site says its levels are calculated "based on market data and various leverage amounts". These tiers are the model's assumptions; they say nothing about the leverage any particular platform offers.
- An approximate liquidation price for each tier. Roughly, a position is liquidated after a move of about 1 ÷ leverage against it: about 10% at 10x, 4% at 25x, 2% at 50x and 1% at 100x. Real liquidation prices sit a little closer to entry, because maintenance margin and fees count too. Our guide to the estimated liquidation price shows the full calculation.
- Stack and colour. The estimates for every tier and every entry level are added up at each price, then shaded.
The same documentation is plain about what the result is: the intensity is a model output that "must not be treated as an actual liquidation amount". That site also publishes three heatmap models, and says they "are calculated separately", so their intensities cannot be compared or added. Two charts of the same market can disagree because they start from different assumptions.
An annotated liquidation heatmap, level by level
Illustrative example, not real market data. A crypto liquidation heatmap for a perpetual contract trading at 60,000. The model has seen open interest added near 60,000 (recently) and near 64,000 (earlier). It uses the four leverage tiers above and, to keep the arithmetic simple, ignores maintenance margin and fees. Each row below is one horizontal band on the chart.
Price level | Side | Positions the model places here | Estimated liquidation value (illustrative) | Shade on the chart |
|---|---|---|---|---|
66,000 | Shorts | 10x shorts opened near 60,000 | $20m | Medium |
62,400 | Shorts | 25x shorts opened near 60,000 | $35m | Bright |
61,200 | Shorts | 50x shorts opened near 60,000 | $15m | Medium |
60,600 | Shorts | 100x shorts opened near 60,000 | $10m | Faint |
60,000 | Current price | |||
59,400 | Longs | 100x longs opened near 60,000 | $12m | Faint |
58,800 | Longs | 50x longs opened near 60,000 | $20m | Medium |
57,600 | Longs | 25x longs opened near 60,000 and 10x longs opened near 64,000 | $70m ($40m + $30m) | Brightest |
54,000 | Longs | 10x longs opened near 60,000 | $25m | Medium |
Four things to read from it.
1. The brightest band is an overlap. 57,600 is 4% below 60,000 and also 10% below 64,000. Two different groups of assumed positions land on the same level, and the model adds them. A bright band means several assumptions agree on a price. It does not mean one large trader sits there.
2. The bands nearest the price are the weakest. The 59,400 and 60,600 bands exist only because the model assumes some positions use 100x. Halve that assumption and they fade; drop it and they disappear. The closer a band is to the price, the more it depends on the highest leverage tier.
3. The imbalance comes from the model. The chart shows $127m of estimated long liquidations below the price and $80m of short liquidations above it. That gap reflects the model's long and short split and the earlier entries near 64,000. It is not a count of real positions.
4. Bands can fade. Depending on the model, a band may dim once price trades through it, on the assumption that those positions were closed. Some of those traders will have added margin or closed early, so the fade is an estimate too.
What the clusters do not predict
A heatmap is useful for seeing where forced closes might pile up. It is easy to read more into it than it holds.
- Direction. A bright band below the price says nothing about whether the price goes down.
- Arrival. A cluster is not a target. Price can stay away from a band, or the band can fade as positions close.
- Timing. Nothing on the chart says when, or if, a level is reached.
- Amounts. The values are model output, not actual liquidations, as the documentation quoted above says.
- What the model cannot see. Stop-loss orders, spot selling, positions on low leverage or cross margin whose liquidation price is far away, and traders who add margin as price moves.
- Real orders. An order book shows limit orders that exist now. A heatmap shows forced closes that might happen, under assumptions.
Research on past crashes points the same way. A 2026 study of seven bitcoin liquidation cascades between 2022 and 2025, including a record USD 19 billion event on 10 October 2025, found that "no variable is event-invariant": the classic early-warning pattern showed up before some cascades and not others, and price gave no such warning before the two sudden tariff-news shocks. The one pattern found in every event with data worked only as a group-level signal, not as an alarm for any single event. The paper is a preprint, not yet peer reviewed.
What a cluster does describe is a risk. If price reaches a large band, forced closes of longs are executed as sells (and of shorts as buys), which can add to a fast move in the same direction. That chain is called a liquidation cascade. It is a reason to know where your own position stands, not a forecast that it will happen. Why visible clusters of stops and liquidations draw attention is covered in our guide to stop hunting.
How to use a liquidation heatmap for your own risk
- Know your liquidation price first. Work it out before you open the position, from your entry, your leverage and your margin.
- Find the nearest bright band on your side. A long position's risk sits in the bands below the price; a short position's sits above. If your liquidation price sits just beyond a bright band, a fast move through that band could reach you too.
- Change what you control. Lower leverage moves your liquidation price further from entry. A stop-loss order closes the position at a level you choose before liquidation does, although in a fast market it can fill at a worse price.
- Compare two models. If two heatmaps disagree on a level, treat that level as weak.
- Check again. Bands move as open interest changes, so a chart from this morning may not match this afternoon.
A heatmap is one input to crypto risk management, not a replacement for position sizing. Trading leveraged products carries a high risk of loss.
FAQ: liquidation heatmaps
What is a liquidation heatmap?
A liquidation heatmap is a chart that estimates the price levels where leveraged futures positions would be forced to close. Price runs up the side, time runs along the bottom, and brighter colours mark more estimated liquidations. It is built from open interest and assumed leverage, not from a list of real positions.
Is a liquidation heatmap accurate?
It is an estimate. The model guesses where positions were opened and which leverage they use, so the bands are only as good as those assumptions. One data site's own documentation says its intensity must not be treated as an actual liquidation amount, and it publishes three models that cannot be compared with each other.
Does price move towards liquidation clusters?
A heatmap cannot tell you that. Clusters show where forced closes might pile up if price gets there, not whether it will. If price does reach a large band, forced closes can add to a fast move, which is a risk to plan for. A 2026 study of seven bitcoin cascades found no sign that warned of each one.
What do the colours on a liquidation heatmap mean?
The colour shows the relative size of estimated liquidations at each price. Dark areas have few or none; bright bands have many. Scales differ between charts and between models, so compare brightness within one chart, not across sites. Bands below the current price are long liquidations, and bands above it are short liquidations.
Conclusion
A liquidation heatmap is a map of estimates: open interest, spread across assumed leverage tiers, stacked at the prices where those positions would be closed. Read the bright bands nearest the price, check which side they sit on, and remember what they cannot show: direction, timing, or real amounts. The most useful level on the chart is your own liquidation price.
Trading leveraged products carries a high risk of loss. See how perpetual futures work before you choose your leverage.
Sources
- CoinGlass API documentation: Pair Liquidation Heatmap (Model 1), for the method ("market data and various leverage amounts") and the model-output caveat
- CoinGlass API documentation: Pair Liquidation Heatmap (Model 2), for the three models being calculated separately
- CoinGlass API documentation: Pair Liquidation Map, for the map's intensity by price level
- Garcia Seuma, R. M.: Where does the criticality live? Early-warning signals are event-heterogeneous across seven crypto-perpetual liquidation cascades, arXiv:2607.27070 (preprint, 29 July 2026)
- Wikipedia: Open interest, for the definition



