Liquidation screener: what it catches and what the alert shows

The liquidation screener looks for futures pairs where the crowd has piled onto one side and is paying funding for it. Two conditions must hold at once: accounts on the long side (or the short side) outnumber the other side several times over, and that crowded side is the one paying. It is a balance of forces a liquidation cascade may come out of, or may not.

In short

The screener looks for futures pairs where accounts are several times heavier on one side and that side is paying funding.
The long-to-short ratio counts accounts, not money: a thousand small longs weigh exactly as much as a thousand large ones.
The score from 0 to 100 is a sum of four observable signs, not a probability of a cascade.
One pair fires at most once every 4 hours. The screener works on futures only.

How the screener decides a side is crowded

The ratio counts accounts, not money

The screener takes the exchange's public statistic on what share of accounts is long and what share is short. A value of 2.6 means 2.6 long accounts for every short account, roughly 72% of participants against 28%. Position size does not enter that number at all: a hundred small longs and a hundred large ones weigh the same. That is why skew is read together with open interest — it shows how much money actually stands behind those positions.

The source is one venue, and the alert names it

Positioning comes from Binance, and from Bybit for pairs Binance does not cover. Values refresh every 5 minutes across several hundred of the most liquid futures pairs; if no fresh value has arrived for more than 15 minutes, the screener stays quiet on that pair rather than working off old numbers. Open interest, funding, and price still come from your own venues — which is why the ratio's source exchange is always printed next to the number.

Skew is not enough — the funding sign has to agree

The crowded side is the one that pays. Positive funding means longs are paying; negative means shorts are. If the skew and the sign of the rate do not agree, the screener does not fire at all:

Skew	Funding	What the screener does
Longs above the threshold	Positive	A “longs squeezed” alert: the crowd is long and paying for it
Longs above the threshold	Negative	Stays quiet: the ones paying are not the ones crowding
Shorts above the threshold	Negative	A “shorts squeezed” alert
Shorts above the threshold	Positive	Stays quiet

What the rate is and why anyone pays it is covered in the article on the funding rate.

What the setup score is built from

The score is a sum of four signs, each with its own ceiling in points. It cannot exceed 100, and no single sign clears a threshold on its own.

Sign	Points	Full score when
Side skew	40	the skew is twice your own threshold
Funding cost	25	the rate is 0.05% or more per 8 hours
Open-interest build	20	OI rose 3% over the hour
Stalling price	15	price has not moved over that same hour

Skew is measured from your threshold, not from parity

This is the main thing to take from the table. Points come not from the ratio itself but from how far it overshoots the bar you set. Exactly at the threshold it is 0 points; at double the threshold it is the full 40. With a 2× threshold a ratio of 2.6 scores 12 points, while with a 1.5× threshold the same ratio scores 29.

Funding is normalised to 8 hours

Venues settle funding on different clocks: some every 8 hours, some every 4, some hourly. So that the same cost of carry scores the same everywhere, the rate is converted to an 8-hour basis for the score. The alert itself still shows the rate exactly as the exchange publishes it — the one you will see in the terminal.

Falling OI scores nothing at all

The open-interest build only counts when it is rising: fresh money is joining the skew. Shrinking OI means the opposite — positions are already closing, so the move is not building, it is happening. That sign scores 0 points, not negative ones.

A price move zeroes the fourth sign

The full 15 points go to a price that is not moving: the crowd has piled in, keeps paying, and gets nothing for it. A move of 1% or more over that same hour zeroes the sign entirely — the setup has already broken up, whichever way price went.

What arrives in the alert: an illustrative example

Take these settings: skew from 2×, score from 60, both directions, open interest from $5m, turnover from $10m. The screener finds a futures pair where, at the moment of the check:

the ratio is 2.6 long accounts per short account (Binance);
the funding rate is +0.06% per 8 hours — longs are paying;
open interest is up 1.8% over the hour;
price has moved −0.2% over that same hour.

Adding up the points: a ratio of 2.6 against a threshold of 2 overshoots the bar by 0.3 — that is 12 out of 40. Funding is above the 0.05% reference, so all 25. The 1.8% OI build out of the 3% needed for the maximum gives 12 out of 20. A 0.2% move out of 1% gives 12 out of 15. That is 61: the threshold of 60 is cleared, and the alert goes out. Set the skew threshold to 3× and the same pair would fail the very first condition, so the screener would never even look at it.

The message will carry: the squeezed side (longs or shorts), the side ratio with the name of the source venue, the funding rate, the nearest stop cluster, a score in the form 61/100, the price on the venue with the largest open interest, the 24-hour range, and links to the exchange and the chart. The screener has no separate alert limit — frequency is held by a 4-hour cooldown per pair.

How the screener helps

Side skew, the funding rate, and the hourly change in open interest would have to be pieced together by hand across three screens — for every pair separately. The liquidation screener does it across every futures pair on 9 exchanges and sends an alert once the skew and the setup score clear your thresholds. The alert carries the side ratio with its source venue, the funding rate, the nearest stop cluster, and the score. No need to pick a coin.

Open the liquidation screener

Why you get too many alerts, or too few

This screener is rare by nature: a side skew is not an every-hour event. Here is what actually governs the flow.

A threshold above 60 demands skew, not just funding

The three signs other than skew add up to 60 points at most. So any threshold above 60 is unreachable without a skew well above your own threshold — and without near-perfect readings on the other three. The default of 75 is a deliberately strict setting: the screener will be silent for weeks on it. For an earlier entry it is usually dropped to 55–65.

Direction drops half the hits

Choosing “longs squeezed only” or “shorts squeezed only” is not a volume dial — it is giving up half the hits. If you are waiting for alerts and none come, check this setting first: it looks harmless and costs more than the thresholds do.

Liquidity floors narrow the list more than you would think

Open interest and turnover are set in millions of dollars and are checked on each venue separately: a pair passes if at least one of your futures exchanges clears both floors. A skew on a thin pair is not a balance of forces but a couple of large orders, so removing those floors entirely is a bad trade — what liquidity is and why it decides things here is covered in the article on liquidity.

Silence can also mean missing data

Positioning is collected across several hundred of the most liquid futures pairs. For a pair outside that list there simply is no ratio, and the screener will not fire on it at any threshold. The same happens when no fresh value has arrived for more than 15 minutes: firing on twenty-minute-old positioning is worse than staying quiet.

What the screener does not do

The score describes the balance of forces
at the moment of the check — nothing more

It does not compute a probability. 85 out of 100 does not mean “fires in 85 cases out of 100”. Behind the number sit four rules with different weights, not outcome statistics, and calling it a forecast would be a lie.
It does not draw a liquidation heatmap. The “nearest stops” figure is the 24-hour price extreme, or a 1.5% offset when that extreme is closer than 1.5% to the current price. A real map needs leverage distribution data, which the service does not have.
It does not see your position. The screener knows neither your leverage, nor your entry, nor your size. What leverage is and how it turns a move into a liquidation is covered in the article on leverage.
It does not promise a cascade. A skew sometimes holds for days and unwinds quietly, with no sharp move. The opposite happens too: the crowd turns out to be right and price goes its way.
It does not name your entry. It shows the pair and the numbers that explain why. Entry, stop, and position size are yours.

Liquidation screener FAQ

What does the liquidation screener catch?

Pairs where one side outweighs the other by at least the factor you set, counted in accounts, and where that crowded side is the one paying funding. Skew alone is not enough: while holding the position costs nothing, the crowd has no reason to close.

Where does the long-to-short ratio come from?

From Binance public data on account shares, and from Bybit for pairs Binance does not cover. Values refresh every 5 minutes across several hundred of the most liquid pairs. The source venue is always named in the alert: positioning comes from one exchange, while open interest, funding, and price are taken from yours.

Is a score of 78 out of 100 a probability of a cascade?

No. It is a weighted sum of four signs: side skew, funding cost, the hourly build in open interest, and a stalling price. There are no outcome statistics behind that number, so it is not a probability and cannot be treated as one.

Why does the screener stay quiet for days?

Usually because the thresholds are high. Three of the four signs add up to 60 points at most, so any threshold above 60 also demands a skew well above your own ratio threshold. The other common reason is the liquidity floors and the chosen direction: picking only long squeezes or only short squeezes drops half the hits.

What does “nearest stop cluster” mean?

The 24-hour price extreme — the level where stops and liquidation orders usually sit. If that extreme is closer than 1.5% to the current price, the screener shows a 1.5% offset instead. It is not a liquidation heatmap: that cannot be computed without leverage distribution data, and we will not invent a number in place of a calculation.

Does the liquidation screener work on spot?

No. Neither open interest nor a funding rate exists on spot, and without them three of the four scoring signs have nothing to work from. The screener only looks at the futures venues on your list.

How often do alerts arrive for one pair?

At most once every 4 hours. A side skew lives a long time — often a day or more — and repeating the same case every half hour would turn it into noise. That is also why this screener has no separate alert limit.

Every field in the form — thresholds, direction, filters — is explained in “Liquidation screener settings”, together with four ready-made setups. The neighbouring price levels screener answers the adjacent question: where exactly the levels sit that stops cluster behind.

Important: side skew is one layer of the picture, not a trading system. Check the alert against the chart, volume, and levels before you open anything.