Liquidity metrics

Liquidity measures the ability to buy/sell an asset without significant impact on price

Liquidity scoring

The maximum score is 10 points, and the minimum score is 1 point. Higher score means higher liquidity, which is good for an asset. The score is calculated based on 3 variables:

  • listing prevalence

  • handy liquidity (in BTC)

  • turnover frequency (for 24h)


Score distribution

Listing prevalence

- 0 points <=0.15

- 1 points 0.15-0.3

- 2 points 0.3-0.6

- 3 points >0.6

Handy liquidity (in BTC)

- 0 points - 0-10 BTC

- 1 point - 10-100 BTC

- 2 points - 100-500 BTC

- 3 points - >500 BTC

Turnover frequency (for 24h)

- 0 points - 0-10,000

- 1 points - 10,000-50,000

- 2 point - 50,000-100,000

- 3 points - >100,000

See the empirical research here.

Liquidity metrics




Bid-ask spread, %

The %-difference between best offer and best bid

More liquid assets have lower bid-ask spreads

10 BTC liquidity

The ratio of best bid-ask spread by 10 BTC (or equivalent) to the best bid-ask spread

Lower ratio means higher liquidity

Handy liquidity

The cumulative volume in an order book at levels remote from the naive mid market price by 0.5% or less

More liquid assets have higher handy liquidity

Volume to marcap ratio

The ratio of a daily trading volume (total by all symbols) to the market capitalisation of an asset

More liquid assets have higher ratio

Number of trades

Number of trades within the latest period

More liquid assets tend to have greater number of trades

Listing prevalence

The function of the number of exchanges where an asset is listed

More liquid assets are listed at a greater number of exchanges

The formulas of the selected metrics are shown below.

Bid-ask spread:

BAS=bestaskbestbidbestaskBAS = \frac{bestask - best bid}{bestask}

10 BTC liquidity:

10BTCliquidity=BAS(x)BAS10 BTCliquidity = \frac{BAS(x)}{BAS}


  • BAS is a bid-ask spread based on best bid and ask

  • BAS(X) is a bid-ask spread by X units of base currency

  • X is 10 BTC or equivalent

Listing prevalence:

L=erf(nN)L = erf(\frac{n}{N})


  • erf is the error function

  • n is the number of exchanges where an asset is listed

  • N = 20