Liquidation Heatmap: What It Shows and What It Doesn't
Lesson 6 of 9 · Advanced
What is a liquidation heatmap?
A liquidation heatmap (or liquidation map) is a chart that uses colour or columns to show at which price levels a large volume of leveraged positions would probably be liquidated. These charts are hugely popular on social media and are often accompanied by lines like "price is going to take the liquidity".
The single most important point of this lesson: a liquidation heatmap is a modelled estimate, not data from the exchange's liquidation engine. Exchanges do not publish the liquidation price of individual users. What you see is the output of a model built on the tool maker's own assumptions.
Separate three kinds of data:
- Liquidation heatmap: an estimate of where liquidations might be.
- Recorded liquidations: positions that actually were liquidated and that exchanges report (although some exchanges limit or aggregate the reporting).
- Order book liquidity heatmap: limit orders that are resting in the book right now; entirely different data, covered in the Crypto Order Flow course.
How are these maps built?
Methods differ between tools, but the general logic is:
- The model tracks changes in OI over time. When OI rises within a price range, it assumes new positions were opened at that price.
- Since the model does not know the leverage of each position, it assumes a set of leverage levels (for example 10x, 25x, 50x and 100x) with assumed weights.
- For each assumed leverage, it calculates an approximate liquidation price and draws it on the chart as a band or colour.
- When price crosses a level or OI falls, the model assumes some of those positions were closed or liquidated and fades or removes the band.
Hypothetical example: if OI grew at $90,000 and the model assumes part of it is long at 25x, an estimated liquidation cluster appears about 4% lower, near $86,000. That cluster is only as real as the "25x" assumption.
What the map does not know
The main limitations that most tutorials leave out:
- Real leverage and margin are unknown: a user may have entered at 5x, or in cross mode with a large balance, which pushes the liquidation price much further away.
- Hedged positions are counted twice: someone holding spot Bitcoin who shorted a perp as a hedge is a "liquidatable short" in the model, though by adding collateral they may never be liquidated. Arbitrage and delta-neutral positions are the same.
- Direction is a guess: rising OI does not say whether the new position was an aggressive long or short; both sides open.
- Stops and manual closes are invisible: many positions are closed by a stop or a manual decision long before the liquidation price.
- Exchange coverage is incomplete: models usually cover only some exchanges, and on-chain platforms or CME may not be included.
For this reason, maps from different tools can give different pictures for the same moment.
The "price magnet" myth
A common claim is that price is drawn like a magnet toward big liquidation clusters. The reasonable core of the idea is that if price reaches a real cluster of liquidations, forced orders can make the move faster. But there are serious problems:
- Selection bias: after any move, you can always find a cluster that price "reached". Clusters that price never reached are forgotten.
- Clusters on both sides: there are usually clusters both above and below price, and the map does not say which comes first.
- An estimated cluster may not exist at all: for the reasons in the previous section.
There is another mechanism too: because these maps are public, some traders deliberately trade toward obvious clusters while others exit just before them. The result is that price behaviour near clusters can go either way: a fast break through, or a reversal before reaching them. The map does not distinguish between the two.
If you use a map, place it next to other data: has OI really built up in that range? Does funding show which side is more crowded? And when price reaches the area, what do volume and delta say? The map only raises the question; live data supplies the answer.
A more balanced reading: the map says that if price reaches an area, there might be extra fuel for the move. It is context, not a price target and not a signal.
Risk warning: your own liquidation price and stop are on other people's maps too. If your stop or liquidation level sits in the same zone everyone can see, in a cascade it may execute with heavy slippage. Highly leveraged positions near obvious clusters are the most fragile. This lesson is educational and does not suggest any price level.
Common mistakes
- Reading a liquidation heatmap as hard exchange data instead of a model's estimate.
- Using map clusters as a price target or entry signal.
- Confusing the liquidation heatmap with an order book liquidity heatmap or with recorded liquidations.
- Ignoring hedged positions that look liquidatable in the model but in reality are not.
- Remembering the times price reached a cluster and forgetting the times it did not.
- Placing your own stop or liquidation level exactly in the clusters everyone can see.
What do volume and data say here?
The best way to judge a liquidation map is to compare it with real data after the fact:
- Recorded liquidations vs the estimated cluster: when price crosses a cluster, check the actual liquidation statistics. If no significant spike was recorded, the cluster was probably overestimated.
- OI drop when crossing a cluster: real liquidations reduce OI. Passing through a "big cluster" with no OI drop is a sign of model error.
- Delta and volume at the cluster: if liquidations occur, you should see a spike in same-direction market orders and high volume; afterwards, see whether absorption happens or the move continues.
Keep a simple journal and record this comparison for a few weeks; you will see for yourself how reliable the map from your chosen tool is.
📊 See real gold futures volume behind these ideas: free order flow course · today's gold volume map
Frequently asked questions
How accurate is a Bitcoin liquidation heatmap?
Its accuracy depends on the model's assumptions: assumed leverage, the exchanges covered, and how closed positions are removed. Because users' real leverage and margin are unknown, treat it as an estimate, not firm data.
Does price always move toward liquidation liquidity?
No. That claim is largely selection bias. Clusters usually exist on both sides of price, and many are never touched. Reaching a real cluster may accelerate a move, but it does not determine where price goes.
What is the difference between a liquidation heatmap and an order book heatmap?
A liquidation heatmap is an estimate of probable liquidation levels derived from OI. An order book heatmap shows actual resting limit orders, which can be cancelled. The two should not be confused.
Education only, not a recommendation to buy or sell; we never give signals. Price examples are illustrative. Test any method on a demo account first, with strict risk management.