Wash trading risk: reading exchange volume with its data-quality flags
Reported exchange volume is the most quoted and least reliable number in crypto. This page publishes the data-quality flags that sit behind it — how many sampled pairs carried an anomaly flag, how many were stale, how many were never rated — so a turnover figure can be read with the caveats that its own source attaches to it.
Exchange-pair quality flags
Anomaly, stale and unrated pair counts for each sampled exchange, from the same pass as the depth figures.
Observed · Observed
| Exchange | Sampled pairs | Anomaly flags | Stale pairs | Unrated pairs |
|---|---|---|---|---|
| Binance | 79 | 10 | 0 | 79 |
| Coinbase | 39 | 10 | 6 | 39 |
| Kraken | 77 | 17 | 2 | 77 |
| OKX | 21 | 1 | 0 | 21 |
| Gate | 64 | 5 | 0 | 64 |
What you can do with it
Qualify a volume figure before using it
A venue whose sample carries several anomaly flags is not a venue whose volume is fake; it is a venue whose reported volume should not be used as a denominator without saying so. That distinction is the entire content of this page.
Compare depth against turnover
The strongest cheap signal is the relationship between the two: heavy reported turnover with thin measured depth is the pattern that has historically distinguished inflated venues from deep ones. Both figures are on this page for the same pass.
Count the unrated pairs
An unrated pair is unknown, not bad. A venue whose sample is largely unrated has not been assessed, and treating that as a clean bill of health is the most common error in this area.
Take the pattern into the pair
The crypto workspace's market-quality views sit beside the chart for a selected pair, so an anomaly at venue level can be followed into the pair that produced it.
How it works
These are upstream flags, not a Vultax accusation
The anomaly, stale and unrated counts come from the data provider that samples each exchange's pairs. Vultax republishes them with the sample size and the observation time. They are not the output of a Vultax detector, and they are not a finding about any exchange's conduct.
Nothing on this page alleges misconduct by any venue or participant. A statistical flag is a reason to check something, and the strongest honest statement that can be made from one is that a figure should be qualified.
Why volume is the number most worth doubting
Reported volume is self-published, weakly standardised and commercially useful to the venue that reports it. Depth is harder to inflate because it has to be quoted, and turnover that is not supported by depth has to come from somewhere. That relationship is the basis of most published research on the subject, and Vultax's own write-up on the topic is linked below the tables.
Sampling is partial by construction
The counts describe the pairs that were fetched for each exchange, not every pair it lists. A venue with a small sample and no flags has not been cleared. The coverage line states this above the table rather than leaving it to be assumed.
What the literature actually found, and what it did not
Published research on this subject has generally worked from statistical fingerprints — the distribution of trade sizes, the roundness of numbers, the relationship between reported turnover and measurable depth — rather than from any direct evidence about who traded with whom. Those methods identify venues whose reported activity does not behave like organic trading. They do not identify a trade as artificial.
That distinction is the whole of the honest position. A venue can look unusual on every statistical test and have an innocent explanation; a venue can look ordinary and be inflating anyway. What a reader can safely take from this page is that a particular volume figure deserves a caveat, not that anyone has done anything wrong.
Vultax's own write-up of the research is linked below the tables, with the papers it draws on and the limits of each.
What each plan gives you here
What a plan changes on this surface.
| What you get | Free | Essential | Premium | Pro |
|---|---|---|---|---|
| Crypto prices, order books and depth | Included | Included | Included | Included |
| Six-domain Vi IQ detail on every pair | Not included | Included | Included | Included |
| The tables on this site, as JSON and CSV | Included | Included | Included | Included |
| Historical lookback | Current views | 3 days | 7 days | 30 days |
Compared with exchange ranking sites
Ranking sites publish a single trust or confidence grade per exchange, which is convenient and hides its own inputs. What is published here is narrower and more boring: the raw flag counts and the sample they came from, so you can form your own view or check somebody else's grade against them.
Competitor details last checked . They describe other products as their operators published them, not as an assessment of them.
The table in full
Each table carries its own source, observation time, coverage and limits, and can be taken away as JSON or CSV. A missing value means the source did not report it; it is never rendered as a zero.
01 Market snapshot
Reported BTC market depth
Reported cost to move the market by 2%, alongside the corresponding BTC pair and reported turnover.
ObservedObserved 5 rowsWithin this table’s source-time freshness window; prices can change between observations.
| Exchange | Pair | +2% depth (USD) | −2% depth (USD) | Reported 24h volume (USD) | Pair status |
|---|---|---|---|---|---|
| Binance | BTC/USDT | 29,326,903 | 16,124,367 | 1,819,121,329 | No flag reported |
| Coinbase | BTC/USD | 28,466,991 | 11,318,985 | 923,008,150 | No flag reported |
| Kraken | XBT/USD | 25,931,133 | 21,546,500 | 332,053,068 | No flag reported |
| OKX | BTC/USDT | 18,098,524 | 8,423,235 | 685,159,118 | No flag reported |
| Gate | BTC/USDT | 21,363,541 | 29,557,545 | 712,858,917 | No flag reported |
Scroll to see all columns.
CoverageCoinGecko tickers for the configured exchanges; the pack selects each venue’s highest-volume BTC USD, USDT or USDC pair.
Depth and turnover are reported by the upstream source, not independently verified executable liquidity.
The upstream pack rounds amounts and may convert missing depth or volume to zero; a reported zero is not proof of no liquidity.
Questions
Does Vultax detect wash trading?
This page publishes upstream data-quality flags and the depth-versus-turnover comparison that qualifies a volume figure. It does not name a trade as artificial and it does not accuse a venue of anything.
What is an anomaly flag?
A marker set by the upstream data source on a sampled exchange pair. Its meaning belongs to that source; the count is republished with the sample size and observation time attached.
Is an unrated pair a bad pair?
No. It is unknown. Treating unknown as clean is the mistake this page exists to make harder.
Can I see which pair triggered a flag?
The published table is at venue level. The crypto workspace is where a specific pair is examined against its own book and trade behaviour.
Which plan do I need?
This page is free. The market-quality views in the crypto workspace sit on Essential and above.
Is this legal or regulatory evidence?
No. It is statistical context republished with its source. It establishes nothing about intent and is not an allegation against any exchange or participant.
Coverage and methodology
- Public tables mirror existing Vultax source packs. They do not include account data, private watchlists or the complete app dataset.
- Source time, coverage and limitations apply to each table. Missing values are unavailable, not zero.
Published research behind this page
Dated write-ups of the same sources. A figure in a study describes the window it was written about, not the snapshot above.
Published estimates put fake or non-economic bitcoin volume at 51% to 95%, and above 70% on unregulated venues. Why the studies disagree, why depth and slippage judge a venue better than volume (Binance: 64.3% of volume, 30.7% of depth, per Kaiko), and what to check yourself.
Order-book feed latency on 16 crypto exchanges: medians from 35.0 ms (Coinbase) to 111.5 ms (BTSE), a 95th percentile of 288–590 ms on every venue, and data in every minute of the day. Plus the coverage check that catches a feed that stays connected and stores almost nothing.
Keep reading
Take this into the workspace
The quality flags here are public. Essential adds the market-quality views beside a pair inside the app.
A free account needs no payment card and starts no subscription. Figures on this page were observed . JSON · Markdown · Schema