A venue that says it lists 1,300 markets is describing its own database, not the number of things you can trade at a useful size. No standard defines the word “market,” so each exchange counts the unit that makes its number largest: every quote currency, every leverage tier, every product line, and in some cases every pair that has ever existed on the platform. The count is an inventory figure, and inventory is not liquidity.

That distinction matters because market count is routinely used as a proxy for depth. A venue advertising 1,300 markets sounds broader than one advertising 300, but the second can carry more executable size in the assets people actually trade. Reconstructing the number is not difficult. It requires separating the four things a market can mean, then applying the same threshold to each venue you are comparing.
- A market is a row in a database. It may be a symbol, a pair, a product, or a single contract expiry, and each definition produces a different total from the same venue.
- Quote currencies multiply the count. One asset priced against six quote currencies can be presented as six markets while all the size sits in one order book.
- Dead pairs stay listed. A pair with zero trades in 24 hours still has a page, a ticker, and often a status that reads as active.
- Leverage tiers are not extra liquidity. Counting 10x, 20x, 50x and 100x as separate markets turns one book into four without adding a single resting order.
Why “market” has no agreed definition
Three units are in common use, and none of them is wrong. A trading symbol is what the API returns. An instrument is what the matching engine runs. A product is what the marketing page sells. The same venue can report all three, and the totals will differ by a factor of five or more.
Nothing forces an exchange to disclose which unit it used. A footnote saying “markets” without a definition is not a measurement, because it cannot be reproduced by anyone reading it. This is the same problem that makes fee comparisons unreliable once the page changes: the label survives, the definition does not.
Four objects behind one word
Start with the symbol. BTCUSDT is a string that identifies a pair on a specific product. Then the instrument: the spot order book for BTC against USDT, with its own matching rules, tick size and fee schedule. Then the product family: a perpetual contract on the same pair, which is a different instrument with different risk. Then the individual contract: a quarterly future with a named expiry, or a single option strike.
One asset can generate dozens of rows under that hierarchy. BTC against USDT produces a spot book, a perpetual book, several dated futures, and an options chain with strikes and expiries. Count every row and Bitcoin alone accounts for more “markets” than most exchanges listed in their first three years.
The page and the API rarely agree
Public endpoints usually carry a status field per symbol: trading, break, halt, or something similar. The marketing page often does not filter on it. Symbols that were delisted last quarter can remain queryable for years, and their pages stay indexed because removing them would break inbound links and search rankings.
That is why two counts of the same venue can differ without either being fabricated. One is counting rows that exist. The other is counting rows that currently accept orders. The gap between them is the venue’s accumulation of history, presented as present-day breadth.
Four ways a count grows without new liquidity
1. Quote-currency multiplication
List one asset against USDT, USDC, FDUSD, BTC, ETH, EUR and TRY and you have seven markets from one asset. In practice, most of the size sits in one or two of those books. The rest are either thin, priced by reference to the main book, or generated automatically by a convertible-quote product that lets users trade across balances without a dedicated order book.
The multiplication is real in the sense that the rows exist and can be traded. It is misleading in the sense that the seventh book adds almost no capacity. If you are estimating how much you can sell without moving the price, counting quote variants overstates depth by the concentration ratio between the main book and the rest.
2. Zombie pairs and retained pages
A zombie pair is listed, reachable, and empty. The page loads, the ticker updates from an index, and the order book shows a spread wide enough to make any trade irrational. It has zero trades in the last 24 hours and often zero in the last 30 days. It remains in the count because nothing in the definition requires trading activity.
Delisted pairs are a variant of the same problem. The venue stops matching, but keeps the page live with a notice, or redirects it, or simply leaves the URL in place. Search engines still index it, and third-party sites that scraped the page weeks ago still show it as available. A count taken from a sitemap rather than an API can include months of retired inventory.
3. Assets presented as markets
The most common substitution is also the easiest to spot: “we list 300 assets” reappears as “1,300 markets” in a later press release. The arithmetic behind it is usually 300 assets multiplied by four or five quote currencies, sometimes with a small allowance for derivatives. No new asset was listed and no new book was funded. The same inventory was counted at a finer granularity.
This is not necessarily dishonest, but it is not informative either. Asset count tells you something about listing policy and due diligence. Market count, computed this way, tells you how many quote currencies the venue supports. Those are different facts, and only one of them is relevant to whether you can exit a position.
4. Leverage tiers and product families
Some venues present each leverage cap as a separate market: BTC/USDT 10x, BTC/USDT 20x, BTC/USDT 50x, BTC/USDT 100x. The underlying book may be shared or separate, but the risk and the margin rules differ, so the rows are defensible as products. They are not defensible as additional liquidity. Four tiers on one pair is one price discovery process with four collateral arrangements.
The same applies to product families. Spot, margin, perpetual, dated futures and options on one asset can be counted as five markets. A venue with 500 derivatives symbols and 800 spot pairs can advertise 1,300 markets while offering meaningful size in perhaps 200 of them. The derivatives share is worth isolating, because derivative liquidity does not help you sell a spot balance.
| Metric as published | What is often being counted | How it inflates | How to check |
|---|---|---|---|
| Total markets | Every symbol across every product, including halted and delisted rows | Adds history and product variants to present-day capacity | Filter the public API for trading status and compare the totals |
| Listed assets | Unique base assets, sometimes including delisted ones | Small number presented as if it were market count | Deduplicate base assets from the symbol list |
| 24h volume | Notional traded across all rows, including wash-like activity | Aggregates thin books into a single impressive figure | Compare per-pair volume against an aggregator and look at the median pair |
| Quote currencies | USDT, USDC, FDUSD, BTC, ETH, fiat pairs | One asset becomes five to seven markets | Count distinct quotes for the top ten assets |
| Leverage tiers | 10x, 20x, 50x, 100x on the same pair | One book counted four times | Check whether the tiers share an order book or a risk engine |
| Delisted pages | Retired symbols still reachable by URL | Adds inventory that cannot be traded | Search the site for withdrawal or delisting notices |
| Derivatives symbols | Perpetuals, dated futures and options | Dominates the total without adding spot depth | Split the count by product line before comparing venues |
A hypothetical: 1,300 markets from 180 assets
Take a hypothetical venue with 180 listed assets. Give it five quote currencies: USDT, USDC, BTC, ETH and one fiat. That produces up to 900 spot pairs before any filtering. Add perpetual contracts on 60 of the larger assets and the total passes 960. Present four leverage tiers on 80 of those perpetuals and the count reaches roughly 1,280, which rounds comfortably to 1,300 in a press release.
Now apply a definition a reader can reproduce. Count only USDT-quoted spot and perpetual markets, one row per asset per product. The number falls to about 240. Require at least $10,000 of daily volume and it falls to roughly 130. Require at least $1 million and it falls to about 45. The venue did not lie at any step. It chose the unit that produced the largest number, and the largest number is the one that got repeated.
A market count measures how many rows a venue has created. Liquidity measures how many of those rows someone will actually trade with you. The two numbers can differ by an order of magnitude and still describe the same platform on the same day.
What you can check yourself in ten minutes
The check does not require data access beyond the venue’s own public endpoints and one aggregator. Run the same six steps on two venues and the comparison becomes usable.
- Pick twenty symbols at random from the published list. Random selection matters; the featured pairs at the top of the page are the ones with liquidity by construction.
- Record the 24-hour volume for each. Count how many show zero or a figure small enough to round to zero. That share is your zombie rate.
- Pull the public symbol list and filter on trading status. The difference between the full list and the tradable list is the retired inventory.
- Compare the active pair count with an aggregator. Aggregators usually filter for pairs that have traded recently, so their total is closer to a usable definition.
- Count quote currencies for the three largest assets. If each asset appears against six quotes but only one book has size, the multiplication is cosmetic.
- Split spot from derivatives. A venue whose count is mostly derivatives is not a deeper spot venue, and the two numbers should never be added when you are judging exit capacity.
What a defensible count looks like
A count becomes checkable when it carries three things: a unit, a threshold and a date. The unit says whether the row is a symbol, a pair or a contract. The threshold says whether zero-volume pairs are included. The date says when the list was last reconciled against the matching engine. Without all three, the number is a claim about inventory rather than a measurement of a market.
A stronger version separates spot from derivatives, publishes the number of pairs with at least a stated daily volume, and keeps a dated delisting log. That last item is the one venues avoid, because a public list of removals makes the gap between the headline count and the live count visible. It also happens to be the most useful page on the site for anyone judging listing quality. The same discipline applies to product totals: adding unlike products into one figure is the mistake that turns a single business line into an apparently larger market.
The resistance is understandable. In the hypothetical above, a $10,000 daily volume floor removes roughly half of the deduplicated rows, and a $1 million floor leaves fewer than one in five. A venue marketing breadth will not volunteer a definition that shrinks its own number by an order of magnitude. That is precisely why the definition has to come from the reader’s side, applied identically to every venue being compared.
Why the definition reaches your cost
Market count feeds decisions that cost money. If you assume 1,300 markets means 1,300 places to exit, you will overestimate the venue’s capacity to absorb a position and underprice slippage. If you route an order to a pair that looks active but trades three times a day, the fill you get is the spread of a market with no competition in it. Breadth on a listing page is not depth at the touch.
There is a record-keeping consequence too. In jurisdictions where every crypto-to-crypto trade is a disposal event, the number of pairs you can trade is the number of taxable combinations you can create, and a venue with hundreds of quote variants makes a clean transaction history harder to reconstruct. Fewer, deeper books are easier to account for than many thin ones.
This is separate from the older question of whether reported volume is real, and from the question of why two ranking sites disagree about the same venue. Volume and liquidity are different measurements, already covered elsewhere on this site, and ranking methodology is its own problem. Market count is a third number with a third definition, and it deserves the same treatment: read the unit, apply a threshold, note the date, and compare like with like. A venue that cannot supply the definition is not necessarily inflating the count. It is simply asking you to take the largest possible interpretation on trust.
0 comments
No comments yet — ask a question and we will answer it here.