Bitcoin drops four
percent in eleven minutes on a Tuesday afternoon with no news attached, and the
move stops almost exactly where it started, falling faster. Traders watching
afterward search for an explanation and usually settle on manipulation. The more
mundane answer involves leverage, forced closures, and a predictable map of
where those closures were waiting to happen.
Liquidation heatmaps
attempt to map that. They estimate the price levels at which leveraged
positions would be force closed, then display those estimates as colored bands
stacked around the current price. Traders who run heatmap trading software from ATAS or
other web-based sources are looking at a model output, not observed market
data, and that distinction shapes everything about how the tool should be used.
The Mechanics Behind the
Picture
Perpetual futures
dominate crypto
volume, and nearly all of that trading happens on leverage. When a position
moves far enough against its holder that the posted margin no longer covers the
loss, the exchange automatically closes it. The trader does not choose the
timing.
That closure is a market
order. A liquidated long becomes forced selling, a liquidated short becomes
forced buying, and neither cares about price.
The liquidation price of
any position follows directly from its entry price and leverage. A long opened
at 60,000 with 10x leverage liquidates somewhere near 54,000, adjusted slightly
for maintenance margin requirements and fees.
Nobody publishes individual positions, so the levels have to be inferred. Providers reconstruct them from data the exchanges do release:
A
Bitcoin
liquidation heatmap trading chart typically shows price on the vertical axis
and time running horizontally, with brightness or color intensity indicating
the estimated density of liquidations at each level.
Brighter
zones mean more estimated leveraged exposure sitting at that price. Should the
market reach the level, the model expects a burst of forced orders in one
direction. Bands above current price generally represent short liquidations,
since shorts lose money when price rises. Bands below represent longs.
Magnetism
is the idea that price gets drawn toward dense liquidation clusters because
closing those positions generates volume that market makers and larger
participants can trade against. Cascading, in turn, occurs when a price trigger
triggers one cluster; the resulting forced orders push the market further. This
mechanism explains why some crypto moves accelerate violently instead of
finding buyers on the way down.
Every
discussion of what a liquidation heatmap is in crypto trading should include a
serious caveat about accuracy. These are estimates based on assumptions that
are not verifiable.
Exchanges
do not publish position-level data. Whatever a heatmap displays comes from
inference, and different providers using different assumptions produce visibly
different maps of the same market at the same moment.
Comparing two vendors on
the same day is an instructive exercise. The broad zones often agree while the
details diverge considerably, which is roughly what you would expect from
independent models attacking an underdetermined problem. Several gaps between
the model and reality account for most of that divergence:
●
Positions closed manually
before ever reaching their liquidation price, which the model still counts.
●
Stop losses placed above the liquidation level
remove exposure the map assumes is still there.
●
Hedged books,
where a trader holds offsetting exposure on another venue and behaves nothing
like the simulation predicts.
Most models assume
isolated margin, where each position carries its own dedicated collateral and a
clean liquidation price. Traders use cross-margin pool collateral across their
entire account, so liquidation depends on the total portfolio, not any single
entry.
A profitable position
elsewhere can push a liquidation price far from where a simple calculation
would place it. Traders who add collateral to a losing position also adjust
their liquidation price, and the model has no visibility into either behavior.
Once enough participants
watch the same levels, they change. Sophisticated traders position around
anticipated liquidations, and market makers adjust quoting accordingly. This
does not make the tool useless, but it does mean the obvious interpretation gets
crowded quickly. Levels visible to everyone lose their edge faster than levels
nobody is watching.