What Trading Volume Tells You and What It Does Not
What volume measures, why spot and derivatives volume differ, what causes spikes, and the evidence that most unregulated exchange volume is fake.
On this page
- What volume actually measures
- The three units volume is reported in
- Spot volume and derivatives volume are different quantities
- What causes a volume spike
- Volume that confirms and volume that diverges
- Why reported exchange volume is often fake
- What volume cannot tell you
- How to use volume without over-reading it
- Conclusion
- Frequently asked questions
- Sources and further reading
Quick read
Volume counts how much changed hands, and almost every conclusion traders draw from it goes further than the number supports. This explainer covers what volume measures, why spot and derivatives volume are different quantities, what actually causes a spike, and why a large share of reported exchange volume has been shown to be fake.
What to remember
- Volume is not buying pressure. Every trade has a buyer and a seller, so total buy volume always equals total sell volume by construction.
- Spot volume and derivatives volume measure different things and cannot be added together. Perpetual futures volume is notional exposure, not assets changing hands.
- A volume spike tells you a lot of activity occurred. It does not tell you the direction, the participant, or whether the move will continue.
- Peer-reviewed testing of 29 exchanges found wash trading averaged over 70% of reported volume on the unregulated venues in the sample.
- The one genuinely useful volume derivative on most exchange APIs is the taker buy share, because it separates aggressive flow from passive absorption.
What volume actually measures
Volume is the total quantity traded during a period. That is the entire definition, and it is narrower than the way volume is usually discussed.
Exchange data makes the narrowness visible. Binance's endpoint returns, for each candle, the base asset volume, the quote asset volume, the number of trades, and the taker buy volumes in both base and quote terms, alongside the four prices. Those are counts of executed quantity. None of them is a measure of pressure, conviction, interest, or intent.
The most important consequence follows immediately: every executed trade has a buyer and a seller for the same quantity. There is no such thing as more buying than selling. When someone says volume shows buyers are in control, they are describing something the number cannot contain.
The three units volume is reported in
The same activity produces three different numbers depending on the unit, and comparisons that mix them are meaningless.
| Unit | What it counts | Comparable across | Common misuse |
|---|---|---|---|
| Base asset volume | Quantity of the asset traded, for example BTC | The same pair over time | Comparing across assets. 1,000 units of two different tokens are not comparable amounts |
| Quote asset volume | Value traded in the quote currency, for example USDT | Different assets and different pairs | Comparing across time without noting that a rising price inflates the figure at constant activity |
| Contracts | Number of derivative contracts traded, converted to a notional value | The same contract over time | Adding it to spot volume, which double counts and mixes leveraged exposure with settled assets |
The quote-volume trap is worth stating plainly, because it produces a recurring false narrative. If an asset doubles in price and the same number of units trade each day, its dollar volume doubles. Headlines describing surging volume during a rally are frequently describing the rally itself, measured twice.
Spot volume and derivatives volume are different quantities
This distinction matters more in crypto than in most markets, because crypto derivatives volume routinely exceeds spot volume by a wide margin, and the two are constantly reported side by side as though they were the same measurement.
Spot volume is assets actually changing ownership. A buyer ends the trade holding the coin and a seller ends it holding the quote currency. The trade requires the full value to be funded.
Derivatives volume, and in crypto that mostly means perpetual futures, is exposure opened and closed. No underlying asset changes hands. A trader at ten times leverage generates ten units of notional volume per unit of capital committed, and can generate that same notional repeatedly within a session by opening and closing.
| Question | Spot volume | Derivatives volume |
|---|---|---|
| Did assets change hands | Yes, in full | No. Only a contract obligation was created or closed |
| Does it require full capital | Yes | No. Margin is a fraction of notional |
| Does it directly affect circulating supply held | Yes, coins move between holders | No, unless the venue hedges into spot |
| Is it a proxy for real demand for the asset | Closer to it, though not clean | Much weaker. It is a proxy for demand for exposure |
| Can the same capital produce it repeatedly in a day | Only by round-tripping the full amount | Yes, and leverage multiplies each round trip |
What this actually tells you
High derivatives volume with flat spot volume is a positioning event, not an accumulation event. Traders are taking exposure rather than acquiring the asset. That configuration tends to be less durable, because leveraged exposure has a carrying cost and a liquidation point, whereas a spot holder has neither.
Adding the two together produces a meaningless total. Aggregators that report a combined figure are summing settled asset transfers with leveraged notional. The result cannot be interpreted, and it inflates whenever leverage rises even if nothing else changes.
The actionable version. When you see a volume surge, establish which market it happened in before drawing any conclusion. A spot-led surge and a perp-led surge have different implications for what happens next, and the perp-led one should send you to open interest and funding rates rather than to a demand narrative.
What causes a volume spike
Volume spikes have a small number of recurring causes, and they are not interchangeable.
| Cause | What is happening | What it implies about continuation |
|---|---|---|
| News or scheduled data | Many participants revalue the asset at once | Depends entirely on whether the news changed the durable picture. The spike itself says nothing |
| Liquidation cascade | Forced closes hit the book in one direction | Frequently exhausts itself. The flow is mechanical and ends when the positions are gone |
| Level break | Resting orders and stops at a price are consumed | The volume is the level being cleared. Continuation depends on what sits beyond it |
| Listing, index or unlock event | A dated, known event forces flow at a specific time | Usually anticipated and priced in advance, so the spike is often the end of the move |
| Wash trading or incentive farming | Volume manufactured for ranking, rewards or appearance | Nothing. It is not information about the market at all |
The last row is not a curiosity. It is common enough that it deserves its own section.
Volume that confirms and volume that diverges
The traditional reading is that volume should expand in the direction of the trend and contract against it. The honest version of this is narrower than the textbook version.
| Price | Volume | Traditional reading | The cautious reading |
|---|---|---|---|
| Rising | Rising | Healthy trend, participation confirming the move | Many participants are transacting near these prices. Says nothing about who is accumulating, since both sides are equal |
| Rising | Falling | Weak rally, likely to fail | Fewer participants are transacting. Could be exhaustion, could be that sellers have simply stopped offering |
| Falling | Rising | Genuine distribution or panic | Often a forced-selling signature. Check liquidations before assuming discretionary sellers |
| Falling | Falling | Drift, sellers exhausted | The most ambiguous of the four. Low activity is compatible with almost any next move |
The one genuinely useful derived number
Most exchange APIs publish a field that is far more informative than raw volume and is almost never used: the taker buy volume. Binance's kline record includes taker buy base asset volume alongside total base asset volume, which lets you separate aggressive flow from passive flow.
A taker is the side that crosses the spread and demands immediate execution. A maker is the side whose resting order was sitting there. Both sides of every trade exist, but only one of them chose to be impatient.
Working it through
Take a candle with total base volume of 1,000 BTC and taker buy base volume of 700 BTC.
- Taker buy volume: 700 BTC, 70% of the total.
- Taker sell volume: 1,000 minus 700 = 300 BTC, 30% of the total.
- Suppose the candle closed roughly where it opened.
Why this configuration occurs. Seventy percent of the executed quantity came from buyers willing to pay the offer, and price went essentially nowhere. That is only possible if resting limit sell orders absorbed the aggression as fast as it arrived. Someone with size was quietly selling into every market buy.
What this actually tells you. Aggression and direction are different things. The conventional read of heavy buying is bullish; the correct read here is that heavy buying met heavier passive supply and failed to move price. That is a genuinely bearish-leaning observation extracted from a number that looks bullish. The reverse configuration, heavy taker selling with a flat close, means passive bids absorbed it.
The actionable version. Compute taker buy divided by total volume for each candle. When that ratio is extreme and price does not respond, mark the price level: passive size is defending it, and that level is worth watching on a retest. When the ratio is extreme and price does respond normally, you have learned much less, because that is simply what aggressive flow usually does. Absorption is the signal; ordinary impact is not.
Why reported exchange volume is often fake
This is the part of the volume story that most tutorials skip, and it changes how much weight the metric can carry.
Exchange volume is self-reported. No regulator audits it on most venues, and until recently the primary consequence of reporting a large number was a higher position on data aggregator rankings, which drives listings, users and fees. The incentive to inflate is direct.
Researchers have tested this systematically rather than assumed it. Cong, Li, Tang and Yang, in Crypto Wash Trading, applied statistical tests to 29 cryptocurrency exchanges, looking for departures from patterns that hold in genuine financial markets: first-significant-digit distributions, size rounding, and transaction tail distributions. Regulated exchanges displayed the expected patterns. The unregulated exchanges did not. The authors quantify on each unregulated exchange in their sample and report that it averaged over 70% of reported volume, describing fabricated volumes in the trillions of dollars annually and documenting how those volumes improved exchange rankings and temporarily distorted prices.
What this actually tells you
The finding is about a sample, not a live measurement. The 70% figure applies to the unregulated exchanges in that study's sample over its data period. It is not a current reading for any specific venue today, and it should never be quoted as one. What generalises is the method and the direction of the incentive, not the number.
Regulated venues behaved differently, which is the useful part. The study's separation of regulated from unregulated venues by statistical signature is the practically actionable result. It means the reliability of a volume figure is predictable from the venue's regulatory status rather than being unknowable.
Wash trading distorts everything downstream of volume. Any indicator that takes volume as an input inherits the contamination: volume-weighted averages, on-balance volume, volume profile, and the volume component of composite sentiment indexes. If the input is manufactured, no amount of processing recovers information.
The actionable version. Source your volume data from the specific venue you actually trade on, and prefer venues subject to a regulatory regime. Treat aggregated cross-exchange volume as a rough order of magnitude rather than a measurement. And when an unfamiliar token shows enormous volume on a venue you have not heard of, the volume is evidence about the venue, not about the token.
What volume cannot tell you
Stated plainly, so that the limits are not buried:
- It cannot tell you direction. Buy volume equals sell volume by definition. Any claim that volume shows more buyers than sellers is arithmetically impossible.
- It cannot tell you who traded. A single large participant and ten thousand small ones produce the same number. The trade count field narrows this slightly, and only slightly.
- It cannot lead price. It is recorded in the same instant as the price it is used to confirm.
- It cannot distinguish real from manufactured activity on its own. That requires either regulatory context or statistical testing of the kind described above.
- It cannot be compared across units or across market types. Base, quote and contract volume are three different quantities, and spot and derivatives volume are not additive.
How to use volume without over-reading it
Steps
Fix the unit and the venue first
Decide whether you are looking at base volume, quote volume or notional contracts, and on which exchange. Most volume disagreements between two traders resolve into one of them looking at a different unit or a different venue.
Separate the spot picture from the derivatives picture
Never combine them. A spot-led surge suggests assets moving between holders; a perp-led surge suggests exposure being taken. Follow a perp-led surge into open interest and funding rather than into a demand story.
Identify what the spike was
News, liquidation cascade, level break, scheduled event, or manufactured volume. These have different implications and only the first two carry information about participant intent.
Compute the taker buy share on the candles that matter
Divide taker buy base volume by total base volume. An extreme ratio combined with no price response identifies a level where passive size is absorbing, which is the one thing volume can tell you that the price chart cannot.
Downgrade volume from signal to context
Use it to judge whether a level was tested seriously and whether liquidity is present for your own size. Do not use it as a directional trigger, because it is coincident and, on some venues, partly fabricated.
Conclusion
Volume is a count of quantity traded, and almost every popular interpretation of it adds something the count does not contain. It cannot show buying pressure, because every trade has both sides in equal measure. It cannot show intent, because a whale and a crowd produce the same number. It cannot lead price, because it is recorded simultaneously with the price it is supposed to confirm.
The distinctions that do matter are structural rather than interpretive. Base, quote and contract volume are three separate quantities and mixing them creates false narratives, most commonly the one where a rally is reported as a volume surge because dollar volume rises mechanically with price. Spot and derivatives volume are not additive at all: one is assets changing hands, the other is leveraged notional exposure that can be regenerated repeatedly by the same capital in a single session.
The reliability problem is more serious than the interpretation problem. Volume on most venues is self-reported and unaudited, and systematic statistical testing across 29 exchanges found wash trading averaging over 70% of reported volume on the unregulated ones in that sample. That contaminates every indicator built on volume as an input, and it means the venue's regulatory status is a better guide to whether a volume figure means anything than the size of the figure itself.
What survives all of this is narrow and genuinely useful. Volume tells you whether a price level was tested seriously and whether there is liquidity for the size you intend to trade. And the taker buy share, published on most exchange APIs and used by almost nobody, tells you where aggressive flow met passive size and failed to move price. That absorption signature is the one thing in this data set that is not already visible on the chart, and it is worth more than every volume-based indicator built on the raw number.
Frequently asked questions
The total quantity traded on a venue during a period, reported either as units of the asset, value in the quote currency, or number of derivative contracts. It is a count of executed quantity, not a measure of pressure or intent, and it covers only that venue's own matched trades.
No. Every trade has a buyer and a seller for the same quantity, so volume carries no directional information by construction. High volume means many participants transacted near those prices, which tells you the level was tested seriously but not which way it resolves.
Spot volume is assets actually changing ownership, funded in full. Futures and perpetual volume is notional exposure opened and closed, where no underlying asset moves and margin is a fraction of notional. The same capital can generate futures volume repeatedly in a session, so the two cannot be added together.
Two reasons. Derivatives notional is frequently reported alongside or combined with spot, which inflates totals with leveraged exposure. And exchange volume is self-reported and unaudited on most venues, with a direct incentive to inflate because aggregator rankings drive listings and users.
Peer-reviewed testing across 29 exchanges found wash trading averaged over 70% of reported volume on the unregulated venues in that sample, while regulated venues showed the statistical patterns expected of genuine markets. That is a finding about a specific sample and period, not a live figure for any venue today.
A sharp rise in traded quantity over a short interval. Common causes are news or scheduled data, liquidation cascades, resting orders being cleared at a broken level, dated events such as listings or unlocks, and manufactured volume. Only the first two carry much information about participant intent.
It describes price moving in one direction while volume moves in the other, most often a rally on falling volume. The description is real but the prediction attached to it is not supported by the data, because volume is coincident with price. Treat it as context for position sizing rather than as a directional trigger.
Taker buy volume is the portion of traded quantity where the buyer crossed the spread for immediate execution rather than resting a limit order. Dividing it by total volume gives the share of aggressive buying. When that share is extreme and price does not move, passive resting size absorbed it, which identifies a defended level.
Treat it as an order of magnitude rather than a measurement. Aggregated figures are sums of numbers that each venue self-reports, so they inherit the reliability of the weakest venue in the sum. Volume from the specific regulated exchange you trade on is far more dependable than a cross-venue total.
Any indicator that takes volume as an input inherits whatever contamination is in that input, so the venue matters more than the formula. If the data is clean, volume is most useful as a liquidity and level-testing check rather than as a directional tool, and the taker buy share adds more than most named volume indicators do.
Sources and further reading
Primary sources:
- Cong, Li, Tang and Yang — Crypto Wash Trading
- Binance — Spot API kline and candlestick data endpoint
Related CoinBeaver articles:
- How to read crypto market signals
- How to read a crypto candlestick chart
- How to read a crypto order book
- Open interest in crypto
- How funding rates work
- How crypto liquidations happen
- Maker and taker fees
This article is educational and is not financial advice. The 1,000 BTC candle used in the taker share example is an illustrative figure chosen to demonstrate the arithmetic, not an observation of any real market. The wash trading estimate is a research finding about the unregulated exchanges in one study's sample over its data period; no current per-venue wash trading percentage is asserted here, because no continuously audited measurement of that quantity exists.
Keep learning
Recommended next reads based on this lesson.
- Altcoin Season: What It Is and How to Measure ItWhy a 90-day rolling index confirms a rotation three quarters of the way through it, the survivorship problem in the top 50, and how to read it honestly.
- Bitcoin Dominance and the Altcoin CycleWhy rising Bitcoin dominance is a half-plane rather than a market condition, what sits in the denominator, and why stablecoin dominance reads cleaner.
- How to Read Crypto Market Signals Without Fooling YourselfThe three families of crypto signals, why a signal with an 80% hit rate is right only 31% of the time it fires, and how to combine signals honestly.
- How Derivatives Expiry Dates Move Crypto Spot PriceHow crypto futures and options expire, how max pain is actually computed, and why convergence toward it is not evidence that price was pulled there.