Decentralised News · 2027 planning series · Sources reviewed 22 November 2026
Market structure · 2027 research edition · Evidence reviewed 22 November 2026
Imagine an exchange announcing $40 billion of trading volume in a market measured at $200 billion. Its share is 20%. A data provider then adds $200 billion of previously uncounted peer activity to the same period. The exchange still has $40 billion. Its share is now 10%. A chart can show a dramatic loss of dominance even though the exchange has not lost a single trade in this deliberately simplified example.
The arithmetic is easy. The editorial mistake happens afterward, when a change in measurement becomes a story about customers abandoning a venue. Before interpreting a market-share chart, establish what could have changed: the venue’s activity, its peers’ activity, the instruments counted, the observation window or the provider’s coverage. Those are different explanations with different consequences for a trader, analyst or business choosing infrastructure.
What Matters
Crypto market share is a fraction built from definitions. Compare the same economic quantity, period, venue universe and counting rules before interpreting a change. A larger denominator can reduce a venue’s share without reducing its activity. Use the DN checker to isolate that arithmetic, then examine depth, execution and data quality separately before choosing a trading venue.
Evidence Block: what the primary documents establish
- Volume has units. Binance’s Spot API distinguishes base-asset volume from quote-asset volume and describes its 24-hour ticker as a rolling window. Those fields cannot be mixed indiscriminately. Binance market-data documentation.
- Publication can lag execution. Deribit’s 10 November 2026 release notes document delayed reporting for qualifying block RFQ trades and separate real-time and delayed volume fields. Deribit release notes.
- Routing and execution are separate measurement layers. DefiLlama keeps aggregator volume separate from underlying DEX volume to avoid counting routed activity twice. DefiLlama data definitions.
- Turnover is not outstanding exposure. CME distinguishes trading volume over a period from contracts remaining open. CME open-interest explanation.
The numerical cases in this article are labelled illustrations. We have not measured current exchange market shares or independently audited a venue’s reported trades.
Start with the question the percentage is supposed to answer
“Which exchange is largest?” is not a sufficiently specified research question. Largest by spot turnover across every listed asset, by Bitcoin options premium, by derivatives notional, or by accessible depth for one order? Each can produce a defensible but different ranking. A venue might dominate a specialised product while representing a modest fraction of a broader market.
A useful share statement names its numerator and denominator in one sentence: this venue accounted for a stated percentage of eligible Bitcoin spot quote turnover across a named set of exchanges during a specified UTC day. That sentence makes the boundaries visible. “Share of crypto trading” conceals them.
Write the measurement contract before collecting the numbers: instruments, venues, currencies, time boundaries, unit conversion, exclusions and revision policy. If a provider changes one of these, preserve the previous definition alongside the new one. The resulting break in a series may be informative, but it should not silently become evidence of competitive movement.
A dollar label does not make two volume figures equivalent
For a hypothetical BTC/USD spot trade, one bitcoin and its dollar purchase value describe the same transaction in different units. Adding both would count it twice. Binance’s documentation makes the distinction explicit through separate volume and quoteVolume fields. Summing base quantities across unrelated tokens is even less meaningful: a thousand units of one token cannot be added to a thousand units of another to create an economically interpretable exchange total.
Converting quote turnover to dollars is a separate calculation. Record the quote currency, conversion price and timing. A stablecoin quote should retain its identity in the raw records; treating its price as exactly one dollar is an assumption, not a property of the word “USD” in a chart label. Where different vendors use different conversion conventions, the apparent disagreement may begin before the market-share calculation.
Options add another distinction. Premium is the amount paid for the option; notional describes a reference exposure using a stated convention. Neither is interchangeable with the other. Deribit’s inverse-options specification quotes premiums in BTC or ETH, while its linear USDC specification uses USDC pricing and instrument-dependent contract multipliers. A count of contracts therefore needs the instrument specification before it can become a comparable value.
Consider a purely illustrative option contract representing ten units of an underlying asset. Ten contracts represent one hundred underlying units. A premium quoted per underlying unit must be multiplied accordingly to recover the cash premium. A separate notional calculation needs its chosen underlying valuation. Keeping these columns separate prevents a visually impressive but dimensionally invalid total.
Outstanding positions cannot be added to daily trading
Volume is a flow over an interval. Open interest is a stock at a point in time. CME’s volume lesson illustrates that closing a position through a trade still contributes to traded volume; volume alone does not establish whether traders are building new exposure. Its open-interest explanation distinguishes outstanding contracts from all transactions during the day.
The practical mistake is a table headed “market size” that adds daily notional turnover to end-of-day open interest. Both may be expressed in dollars, yet the sum has no clear economic meaning. Keep a turnover-share table and an open-interest-share table separate, with their own conventions. Even open-interest comparisons require checking how each source represents the two sides of a matched position.
A high-turnover venue may serve frequent position adjustments. A venue with substantial open positions may have relatively little trading that day. These are useful hypotheses about activity, not a permission to infer customer counts, profitability or loyalty from either number. Those claims require additional observations.
The denominator has a membership list
Return to the fictional $40 billion venue. The first universe contains $200 billion, including that venue. The expanded universe contains $400 billion. Its reported share falls by ten percentage points, from 20% to 10%; the relative decline in the percentage is 50%. Neither description means the venue lost $20 billion of business.
This is why “global” requires evidence. A total assembled from the exchanges a vendor tracks is a tracked-universe total. It may be highly useful, but missing venues do not become zero-volume venues. The provider may omit a region, lack a working feed, exclude certain instruments or decide that a market fails its quality rules. Document those choices before treating the denominator as the entire industry.
For changes over time, produce two views when the data permits. A fixed panel compares the same eligible venues in both periods. An expanded panel incorporates new coverage and labels the coverage change. Neither view is automatically superior: the fixed panel helps examine activity within comparable coverage, while the expanded panel may better represent the presently observed ecosystem.
Do not fabricate a fixed panel by backfilling an absent venue with zeros. “Not observed” and “observed with no trades” need separate states. Also maintain stable venue identifiers. A rebrand, regional entity or newly separated product should not be counted as a new competitor until the mapping shows that it represents distinct activity.
When the 24-hour label hides different clocks
A rolling window ending now and a calendar day ending at midnight can contain different trades while both being described casually as daily volume. Binance’s 24-hour ticker documentation specifies a rolling window and supplies opening and closing times. Preserve those boundaries instead of discarding them when importing a convenient headline number.
Publication time introduces another clock. Deribit’s 10 November 2026 notes describe an eight-hour default reporting delay for qualifying block RFQ trades. Original execution timestamps and sequence numbers are retained, so reports can arrive out of order and later historical queries can include additional trades. The documented 24-hour statistics combine real-time volume with a delayed component covering an offset execution window. This is a vendor-described reporting design, not a DN finding that all Deribit trading is delayed.
An analyst comparing two downloads must therefore ask whether the values refer to execution time, publication time or a vendor-defined mixture. Re-running yesterday’s query can legitimately change yesterday’s total when late records arrive. Store extraction time and source version alongside the economic period, and set a policy for provisional and final observations.
For a daily chart, show whether late records are reassigned to their execution day or retained on their reporting day. Either approach can answer a coherent question if disclosed. Alternating between the two produces a series whose movements cannot be interpreted consistently.
One trade can appear at several layers of the same stack
A trader can use an aggregator that routes an order to underlying liquidity venues. The interface and the execution venue are both involved, but they do not represent two independent customer trades. DefiLlama’s definitions explicitly separate aggregator volume from DEX volume and count an aggregator user’s overall trade rather than adding route legs. It also specifies single-side counting for DEX swaps and taker-side counting for perpetual volume.
The same measurement discipline applies when comparing brands, front ends, protocols and chains. First decide which layer owns the observation in the intended chart. A distribution-share chart may attribute activity to the interface. An execution-share chart may attribute it to the destination. Adding those two charts together creates overlap rather than a larger market.
Build an attribution table with the original trade identifier, execution venue and routing layer where the data supports it. Where identifiers cannot be reconciled, disclose that the series cannot safely be combined. On-chain visibility does not by itself solve a classification error: perfectly observable records can still be counted under the wrong definition.
Reported volume is not a certificate of execution quality
Once the fraction is correctly assembled, it still answers only its stated question. A large share of eligible turnover does not establish that a particular order will execute cheaply, that withdrawals will work or that an exchange has adequate resources. Those are separate decisions requiring different evidence.
CoinGecko’s Trust Score methodology combines several dimensions, including liquidity and operational or regulatory information, rather than ranking exchanges solely by self-reported volume. It is a methodology-specific composite. A score should not be relabelled as market share, a solvency finding or permission to use a service in a particular jurisdiction.
For an execution decision, a more relevant next step may be market depth near the prevailing price for the actual pair and intended size. Kaiko’s depth-comparison example measures available bids and asks within a defined distance of the mid-price and shows that depth varies across exchanges and time. That remains an observable snapshot, not a promise that the displayed orders will remain available.
Use turnover to understand participation within the defined market. Use depth, spread and realised execution evidence to investigate trading conditions. Combining the questions into one ranking creates unwarranted certainty, even if every underlying number is accurate.
DN Market-Share Denominator Checker
This tool isolates one mechanism: adding eligible peer activity to the denominator for the same period while holding the target venue’s volume unchanged. Enter the target’s volume, the original total including that venue, and additional peer volume not already counted. All three inputs must use one selected unit, the same period and compatible counting rules.
DN Market-Share Denominator Checker
Hold a venue’s volume fixed and add eligible peer coverage for the same UTC interval. See how the denominator changes the percentage without inferring a change in customer activity.
This instrument runs locally in your browser. It does not connect to accounts, fetch evidence or submit instructions. Use aggregate values and non-sensitive references. Downloaded records contain your inputs.
Machine-readable illustrative record
An illustrative record of the example assumptions, not an observed transaction or provider test. A null calculation timestamp identifies this static example.
{
"schema_version": "1.0.0",
"article_id": "N008",
"evidence_basis": "illustrative",
"source_review_date": "2026-11-22",
"computed_at": null,
"inputs": {
"venue_volume": "40",
"original_total": "200",
"added_peers": "200",
"unit": "usd_quote_turnover",
"period_start_utc": "2026-11-20T00:00",
"period_end_utc": "2026-11-21T00:00",
"same_period": "yes",
"compatible_basis": "yes",
"non_overlapping": "yes"
},
"status": "Valid coverage sensitivity: target activity held unchanged",
"outputs": {
"tool_id": "N008",
"version": "1.0.0",
"venue_volume": 40,
"original_total": 200,
"added_peers": 200,
"expanded_total": 400,
"unit": "usd_quote_turnover",
"period_start_utc": "2026-11-20T00:00:00.000Z",
"period_end_exclusive_utc": "2026-11-21T00:00:00.000Z",
"same_period": "yes",
"compatible_basis": "yes",
"non_overlapping": "yes",
"original_share_pct": 20,
"expanded_share_pct": 10,
"change_percentage_points": -10,
"relative_share_change_pct": -50
}
}The default illustrative inputs are 40, 200 and 200. They return an original share of 20%, an expanded-universe share of 10% and a change of minus ten percentage points. Adding no peer volume leaves the share unchanged. A zero-volume venue remains at zero; the relative percentage change in a zero starting share is undefined.
The checker refuses missing amounts, a zero original denominator, negative values, a target larger than its original total, or an unconfirmed compatibility statement. It does not fetch volume, validate a vendor or convert premium into notional. Its export records the assumptions so the calculation can be reviewed alongside the underlying data.
The result supports a precise sentence: “Under the expanded coverage, the same target volume represents this smaller fraction.” It cannot establish why real customers traded elsewhere, whether a venue is gaining users or whether a reported trade was economically genuine.
DN Alpha Thesis: publish a bridge, not just a ranking
The strongest market-share analysis explains how the old percentage became the new one. Separate changes in target activity, peer activity, coverage, instrument classification, conversion and reporting policy. That bridge gives a reader something more useful than a league table: an account of which part of the story is actually supported.
Where a clean decomposition is impossible, say which effects remain entangled. A venue entering the dataset while changing its product mix cannot be reduced to one confident competitive explanation. Preserve the uncertainty and identify the additional records needed to resolve it.
For DN’s broader market-structure context, read How Crypto Derivatives Drive Volume, Depth and Dominance. For sourcing the underlying records, the historical crypto tick-data guide addresses data-provider selection. This article’s narrower job is to make a market-share claim reproducible before anyone interprets its competitive meaning.
Methodology and limitations
This is the 2027 research edition, with primary documentation reviewed on 22 November 2026. It compares definitions and demonstrates arithmetic; it does not rank current venues or claim a proprietary trade-level investigation. Exchange specifications and vendor methodologies can change, so record the version used for any subsequent measurement.
The checker assumes the additional peer volume is eligible, non-overlapping and measured consistently. These are declarations by the reader, not facts authenticated by software. A mathematically valid fraction can still rest on poor inputs. When coverage, conversions or reporting times cannot be established, retain the unresolved fields and withhold a competitive conclusion.
Frequently asked questions
Can an exchange lose market share while its own volume rises?
Yes. Its share falls if the comparable denominator grows faster than its own volume. A separate possibility is expanded data coverage. Distinguish actual growth among consistently observed peers from a provider adding previously uncounted activity; the percentage alone does not separate those explanations.
Are options premium and options notional interchangeable?
No. Premium describes the price paid for an option, while notional describes a reference exposure under a stated convention. Contract multipliers and valuation rules matter. Compare premium with premium or consistently defined notional with notional, and label the measure.
Is a 24-hour volume number the same as a UTC daily total?
Not necessarily. A rolling window can end at the moment of extraction, while a UTC day has fixed midnight boundaries. Delayed reporting can add another difference. Preserve both economic-period boundaries and the time at which the data was retrieved.
Why should aggregator volume be kept separate from execution volume?
An aggregator can route the same user trade into an underlying venue. Adding the aggregator total to the execution total can count that activity twice. Choose whether the chart measures distribution or execution, and document the attribution rule.
Does higher volume guarantee a better fill?
No. A volume share describes past activity within its measurement universe. Investigate the actual trading pair, spread, available depth, intended order size and execution evidence separately. Even an observed order-book snapshot can change before an order reaches the market.
What does the DN denominator checker actually prove?
It proves the arithmetic consequence of adding a declared amount of eligible peer activity while holding target activity constant. It does not authenticate the data or infer customer behaviour. Its example is hypothetical, and its compatibility checks depend on accurate reader declarations.
How should missing exchange data be handled?
Keep missing observations distinct from measured zero activity. Report the coverage gap, preserve a consistent panel where useful and show expanded coverage separately. A provider outage or an untracked venue should not silently become evidence that no trading occurred.
Primary sources and review record
- Binance Spot REST market-data documentation. Base versus quote volume; rolling 24-hour ticker and window timestamps. Reviewed 2026-11-22.
- Deribit release notes, 10 November 2026. Qualifying block RFQ delayed reporting, preserved execution timestamps and volume_realtime/volume_delayed components. Vendor-described behaviour, not independently tested. Reviewed 2026-11-22.
- DefiLlama data definitions. Single-side DEX volume, taker-side perpetual volume and separation of aggregator and execution layers. Reviewed 2026-11-22.
- CME: Open interest. Outstanding contracts versus traded volume over an interval. Reviewed 2026-11-22.
- CME: What is volume?. Closing transactions still contribute to volume; volume alone does not identify opening exposure. Reviewed 2026-11-22.
- Deribit inverse options specification. Inverse options premium quoted in BTC/ETH and documented contract units. Reviewed 2026-11-22.
- Deribit linear USDC options specification. USDC pricing and instrument-dependent contract multipliers. Reviewed 2026-11-22.
- CoinGecko Trust Score methodology. Composite exchange scoring across liquidity and other dimensions rather than raw reported volume alone. Reviewed 2026-11-22.
- Kaiko: Compare market depth between exchanges. Depth within a defined distance of mid-price; variation across venues and time. Reviewed 2026-11-22.