Liquidity and market structure
Turnover is observable. Depth is not. Most crypto liquidity analysis quietly treats the first as if it were the second.
Liquidity is usually defined as the ability to trade size without moving the price. That definition refers to a counterfactual — what would happen to the price if you traded — and counterfactuals are not in any data feed. Every practical liquidity measure is therefore a proxy, and the useful question is which proxy fails in which direction.
What is actually observed
| Quantity | Status | Reliability |
|---|---|---|
| Reported 24h volume | Observed, self-reported by venues | Varies sharply by venue |
| Market capitalisation | Derived from a supply estimate | Depends on provider methodology |
| Turnover (volume / market cap) | Derived from the two above | Inherits both weaknesses |
| Realised volatility | Observed from price history | Solid, window-dependent |
| Order book depth | Observable per venue, not aggregated here | Fragmented, changes in milliseconds |
| Price impact | Not observable without trading | Modelled only |
| Liquidation zones | Modelled from assumed leverage | Estimate, not measurement |
Moonboard's LMR is built entirely from the first three rows. It is a turnover ranking, and the page says so.
What turnover does well
Turnover — volume divided by market capitalisation — is the most useful single liquidity proxy available from public data, for one reason: it removes the size effect. Raw volume rankings simply reproduce market cap rankings. Turnover asks what fraction of an asset's value changed hands, which is comparable between a two-hundred-million-dollar asset and a twenty-billion-dollar one.
It is a genuine measurement of trading intensity, and at the low end it is informative in an unambiguous way: an asset turning over 0.1 % of its market cap per day cannot absorb a large position quickly, and no modelling is required to say so.
Where turnover fails
It measures completed trades, not available depth
Two assets with identical turnover can have entirely different order books. One might have deep two-sided liquidity absorbing a steady flow; the other might have a thin book where the same volume arrives as a handful of trades that each move the price several percent. Turnover cannot distinguish them, and the second case is riskier despite looking identical.
High turnover on a thin book is closer to a warning than a reassurance — it indicates price is moving on modest flow.
Reported volume is not verified
Volume figures come from venues with varying incentives and reporting standards. Wash trading — trading with oneself to manufacture apparent activity — inflates the numerator directly, and it is most prevalent exactly where turnover is most likely to be examined: smaller assets on smaller venues seeking visibility.
Moonboard applies no wash-trading filter, and LMR cannot distinguish genuine turnover from manufactured turnover. An unusually high turnover ranking for an otherwise obscure asset warrants checking which venues the volume came from before it is treated as a finding.
Market cap depends on a supply estimate
Circulating supply is a judgement made by data providers about which tokens should count — excluding locked, burned, treasury-held or unvested balances. Providers reach different conclusions, and revise them. A supply revision changes turnover with no change in trading whatsoever.
A day is a short window
A listing, an unlock or a liquidation cascade can put an asset in the top turnover decile for a single day. Turnover describes yesterday. Whether it describes a regime requires looking at the series, not the value.
Observed versus modelled: liquidation zones
This distinction deserves its own section because it is where crypto analytics most often blurs the line, and Moonboard's application includes a liquidation-zone visualisation.
A liquidation heatmap is a model output, not a measurement.
Producing one requires assuming:
- a distribution of leverage across open positions,
- a distribution of entry prices,
- the margin and liquidation rules in force,
- and that positions across venues can be aggregated meaningfully.
None of these is directly observable across the market. The output is a plausible reconstruction under stated assumptions, and its resolution should not be mistaken for precision — a rendering with sharp bands looks far more certain than the inputs justify.
Phrases such as "short squeeze imminent" or "price is forced toward the cluster" state a modelled scenario as a fact about the future. The defensible formulation is that a model, under its assumptions, places estimated liquidation density in a region — a condition worth being aware of, not an event that has been established.
Price impact scales non-linearly
A practical point often lost when liquidity is reduced to a single number: impact does not scale linearly with size. Doubling an order can more than double its price impact, because it consumes progressively worse levels of a book whose depth thins with distance from the mid.
This means a turnover ranking cannot be converted into an executable size, and the gap widens for exactly the small, thin assets where turnover rankings look most exciting.
Volatility and liquidity are related but distinct
Illiquid assets tend to be volatile, because a given flow moves the price further. But high volatility does not imply illiquidity — a deep market can be volatile when information genuinely arrives.
Moonboard treats them as separate inputs throughout, and SDE uses volatility explicitly as its second term. Note there that volatility is unsigned: it does not distinguish a sharp rally from a sharp fall.
Open questions
- How much of reported volume for small-cap assets survives a plausible wash-trading filter, and how much would turnover rankings change if one were applied?
- Does turnover rank predict realised execution cost well enough to be used as a risk input, or only as a screen?
- How stable is turnover ranking over time — is an asset in the top decile today likely to be there next week?
- Can venue-level depth be aggregated into something more informative than turnover without introducing more error than it removes?
Related
- LMR methodologyHow turnover becomes a ranked metric
- Volume-to-market-cap ratioThe underlying calculation
- Market liquidityThe concept, from first principles
- VolatilityThe related but distinct measure