24h turnover screener: how it catches volume spikes

Every minute the 24h turnover screener measures how much money goes through every pair across 10 exchanges and sends an alert when trading activity changes sharply. It covers spot and futures — you choose the market. The key point up front: the screener does not look at the volume bars under the candles, it counts dollar turnover. What the metric itself is, is covered in “What trading volume is and how it differs from turnover”; this article is about what the screener counts as an event and why you may get too many alerts or too few.

In short

Three independent scenarios: the change in 24h turnover over an interval, turnover in a single minute in dollars, and a spike against the coin’s own average.
This is not candle volume. The screener never uses the bars under the chart — it counts dollar turnover from exchange data.
Spikes and per-minute turnover can also go through the wash-volume filter; it never touches the percentage-change scenario.
One move, one alert: the condition has to hold now and not have held a minute earlier.

SOL/USDT Turnover +142.50%

Bybit Futures

Volume confirmed by open-interest change Volume spike vs rolling average ×3.2 of average (baseline $8.4M over 20 min) Volume Surged by 142.50% ($18.2M) Price Increased by 3.10% Price $164.8 24h $152.4 — $168.2

Breakdown

After past alerts 4 of 10

Price mostly reversed: typically −1.1% within 1h (alerts over 30 d)

Alerts on this coin today 1

4 this week

24h turnover $18.2M

Medium liquidity

Volume confirmed

Over 15 min: Price and OI rise. Check OI units and the chart; these readings do not identify who is opening positions.

All indicators (15)

Longs / shorts: 1.24 More longs than shorts, which is normal for the market
Funding, Bybit · futures: +0.0100% Normal rate, no skew
Open interest (15 min): +2.80% Open interest increased; account for its units and the price change
OI changes / turnover: 0.22 OI changes are moderate relative to estimated turnover
To 24h high: −2.02%

1h RSI: 61 Neutral zone: neither overheated nor sold off
From the 200-day average: +5.30% Above the 200-day average: long-term trend is up
From the 50-hour average: +4.10% Above the 50-hour average: short-term trend is up
Daily RSI: 58 Neutral zone: neither overheated nor sold off
Volume vs average: ×3.20 Volume is 3.2× the usual: a lot of money is in the move
Indicators from the chart of: Bybit · futures

Coins moving the same way (15 min): 44% The whole market is moving, the coin follows it

Korean price vs world: +0.60% Price in Korea is above the world price: local buyers are more active
Gap change over 15 min: +0.2 pp The gap with the world price is growing
Korean trading, 24h: $12.1M Enough trading in Korea to trust the gap with the world price

What exactly counts as an event

There are three scenarios and they are independent: any enabled one is enough for an alert. They measure different things, so their thresholds mean different things too.

Scenario	When it fires
Turnover change	The pair’s 24h turnover changed by the set percentage over the chosen interval — anywhere from 1 minute to 4 hours. The threshold is signed: “+30%” is activity rising, “−30%” is activity fading.
Turnover per minute	More than the set dollar amount went through the pair in the last minute.
Spike against the average	The last minute’s turnover is N times the average per-minute turnover of that same coin over the previous 20 minutes.

Percentages here are measured against 24h turnover

The first scenario does not compare candle to candle: it compares 24h turnover with 24h turnover N minutes ago. That turnover is a rolling window: every minute the newest minute of trading is added and the same minute from a day ago drops out. Moving a number like that by tens of percent takes an inflow comparable to a full day’s, which is why thresholds here are set in tens of percent, not in multiples.

The spike compares a coin with itself

Per-minute turnover is derived from the difference in 24h turnover between neighbouring minutes, multiplied by price. The average is taken over the last 20 minutes, and if fewer than ten of those minutes are valid no alert is sent: on a fresh pair or in thin liquidity such an average means nothing. Because each coin is compared with itself, one multiplier is equally meaningful on BTC and on a small altcoin.

The wash-volume filter: what it checks and what it cannot do

Some venues report more turnover than really trades. The filter scores a spike and decides whether it reaches you — per venue, not per coin.

What it measures

The filter builds a score from three observable signs over the 15 minutes around the spike: what share of turnover turned into open positions, whether price moved over the same window, and the ratio of the venue’s open interest to its own 24h turnover. A real flow usually moves price and leaves a trace in positions; churning your own liquidity does neither.

What you configure

Two things. Strictness: the soft mode cuts only obvious wash trading, the strict one also cuts borderline cases. And a separate switch for whether suspicious spikes stay in the feed with a badge rather than being hidden outright. The verdict travels into the alert as a line: volume confirmed by open-interest change — or volume in question, because open interest barely moved.

It does not check everything

The filter applies to two scenarios only — the spike and per-minute turnover. It never touches a percentage change of turnover over 30 or 60 minutes: such a change is not a signature of wash trading, and a 15-minute check window says nothing about it.

When there is nothing to check, the spike goes through

All three signs rest on open interest, and spot has none; a freshly listed pair has too little history. In those cases the verdict is “could not verify”, and the spike reaches you with no badge: dropping the unverified would mean staying silent about the newest moves. We claim no accuracy figure for the filter: there is no labelled sample on which it could be measured.

What the alert contains

The alert opens with the pair and the interval — for example “ARB/USDT | 30-Minute Change Alert”. If you switched on the alert cap, a counter of fires for that coin over the chosen period follows. Then comes a block per exchange where the condition fired, and what is inside depends on which scenario fired:

The venue’s funding rate, if this is a futures pair.
The wash-filter verdict, when the check succeeded.
Volume spike — how many times above average, and what that 20-minute average was.
Per-minute volume — the dollar amount over the last minute.
Turnover change in percent and in dollars, plus the price change over the same interval — these lines belong to the percentage scenario.

Below that: the 24-hour price range, the longs-to-shorts ratio (when fresh data for the pair exists), and the links you picked in the settings — the exchange, the chart, the coin’s page on the site.

How the screener helps

By hand you can spot a spike only on the coin whose chart is open. Every minute the screener measures per-minute turnover on every pair across 10 exchanges, compares it with that same coin’s average, and sends a push or Telegram alert — with the size of the spike, the dollar amount, and the venue it happened on. No need to pick a coin in advance.

Open the 24h turnover screener

Why you get too many alerts or too few

First, what the screener does by itself. One spike will not repeat. Every scenario fires on the edge: the condition has to hold now and not have held a minute earlier. Without that, a “turnover three times the norm” condition that holds for an hour would send an alert every minute. The percentage change and the spike carry one more rule: when the previous minute’s data is missing, no alert goes out — “nothing to check against” is not the same as “it was below the threshold”.

Too many usually means a multiplier with no liquidity floor

On a coin whose normal minute is two hundred dollars, a five-fold spike is one thousand dollars, and pairs like that fire constantly. A floor on 24h turnover and on open interest, both in millions of dollars, removes them at once. The second lever is the pause between alerts on one coin: the first busy minute is usually followed by several more.

Too few usually means a candle-sized percentage

A 500% threshold in the first scenario will almost never fire: percentages there are measured against 24h turnover, not against a one-minute candle. If what you want is “a candle with several times the volume of its neighbours”, that is the “Spike against the average” scenario — it is built as a comparison with the recent norm.

Dead silent? Check the conditions and the market filter

A screener that is switched on with an empty interval and no threshold never fires, while from the outside it looks like it is working: the scenarios block in the form shows a counter of active conditions. The second common reason is a market filter left on one market while the coin trades on the other. Every field is explained in “24h turnover screener settings”.

What this screener does not do

The screener answers one question: where trading activity changed and by how much. Everything else is on you.

It does not say whether that is buying or selling

Turnover is the sum of all trades, and every trade has two sides. Direction comes only from the price line next to it. High turnover while price stands still means buying and selling balanced each other out.

It does not compare candles on a chart

The volume bars under the candles are not part of the calculation at all. The closest thing in spirit is the spike-against-average scenario, and even that counts dollar turnover per minute, not a candle’s volume in coins.

It does not add venues into one number

Each venue is checked on its own, and only those where the condition fired make it into the alert. Wash turnover on one exchange does not kill an honest spike on the same coin elsewhere — the filter cuts the venue, not the coin. You can compare venues on the 24h turnover page.

It does not name an entry, a stop, or a position size

Turnover records trades that have already happened. The alert carries measured numbers only: what changed, by how much, and on which venue. What is behind it and whether to enter is your call.

An illustrative example: how it looks in the settings

Take an illustrative setup for finding sudden activity on liquid altcoins. The numbers are here to illustrate, not taken from the market.

Spike against the average: 5×.
Turnover per minute: from $0.5m.
Minimum 24h turnover: $50m.
Wash-volume filter: soft mode.
Pause between alerts on one coin: 30 minutes.

What that setup catches

A pair whose normal minute is around $120k of turnover, while the current minute did $780k. That is 6.5× its own average and above the per-minute threshold at the same time, so both lines appear in the alert: the size of the spike with the average next to it, and the dollar amount for the minute. The filter checks whether that turnover left a trace in open interest and attaches its verdict. While the activity holds, there will be no repeats.

What that setup misses

A coin with $9m of daily turnover — it does not clear the floor, however loud the multiplier is. It also misses a slow build-up: activity that grows over hours may never reach 5× within one minute. For that case you switch on the first scenario — the change in 24h turnover over 1 to 4 hours.

The threshold is signed. “−40% in an hour” is a scenario of its own: activity on the coin is fading, and the price move now runs on less money than it did yesterday.

24h turnover screener FAQ

What does the 24h turnover screener do?

Every minute it measures how much money goes through every pair across 10 exchanges and sends a push or Telegram alert when trading activity changes sharply. There are three scenarios: the change in 24h turnover over an interval, turnover in a single minute in dollars, and a spike relative to the coin’s own average.

Is this the volume shown under the candles?

No. The screener never uses the volume bars under the candles. It counts dollar turnover: the 24-hour figure as a rolling window of the last day, and the per-minute figure as the difference in 24h turnover between neighbouring minutes, multiplied by price.

What multiplier should I set for a volume spike?

The multiplier compares a coin with itself: the last minute’s turnover against the average per-minute turnover of the previous 20 minutes. That is why one value works on BTC and on a small altcoin alike. The higher the multiplier, the rarer the alerts; always pair it with a floor on 24h turnover.

Does the screener work on spot?

Yes. The market filter lets you watch spot only, futures only, or both. Spot pairs carry no funding line in the alert, and the wash-volume filter cannot reach a verdict on spot: all of its signs rest on open interest, which spot does not have.

How does the wash-volume filter work?

It scores a spike on three observable signs over 15 minutes: what share of turnover turned into open positions, whether price moved, and the ratio of the venue’s open interest to its 24h turnover. Soft mode cuts only obvious wash trading, strict mode also cuts borderline cases. The filter applies to the spike and per-minute scenarios, not to the percentage change.

Why does one spike produce only one alert?

The screener fires on the edge: the condition has to hold now and not have held a minute earlier. Otherwise a spike lasting an hour would send an alert every minute. The next alert for that coin arrives when the condition fires again and the pause you set has passed.

Can I get alerts when turnover falls?

Yes. The first scenario’s threshold is signed: “−40%” means an alert arrives when 24h turnover drops 40% or more over the interval. That is not selling, it is activity fading: less is trading now than in the same minutes a day ago.

Every form field, with example values, is covered in “24h turnover screener settings”. How to tell market interest from one-off activity is in “24h turnover screener use cases”. Next to it sits the open interest screener, which answers not “how much traded” but “how many positions stayed open”.