ETH: Spot Volume Turnover of Book Depth by Venue
Included with Pro.
How many times each venue's reported 24-hour ETH spot volume turned over the resting liquidity within 1% of the mid price on its order book.
Why use it. Reported volume is hard to judge on its own. Measured against the liquidity that actually sits in the book, a very large figure stands out: real trading depends on resting orders, while trades that only circulate do not need them. How to read it. Compare a venue with the others and with its own history. High readings against peers deserve a closer look but are not by themselves a verdict on a venue: a tight, very active market can also turn its book over many times. The summary value is all venues together (total reported volume / total depth).
Calculation. Reported trailing 24-hour volume in US dollars at the end of the period / the mean of bids plus asks within 1% of the mid over the period, for the venues whose book was complete to 1% on both sides. Source: our own live order books of the public spot markets of Binance, Bybit, Coinbase, Kraken and OKX (each coin's USDT market on Binance, Bybit and OKX and its US-dollar market on Coinbase and Kraken). Coinbase, Kraken and OKX since 23 October 2026. Each book is kept from the venue's public feeds: Binance's from a REST snapshot of its best 5,000 or 1,000 levels a side, brought up to date with every 100-millisecond change and the venue's update sequence checked on each; Bybit's from its 1,000-level book stream (a snapshot, then numbered changes); Coinbase's from its public book stream, the whole book as a snapshot and then every change of a price level, with every message numbered and the numbers checked; Kraken's from its 1,000-level book stream, every change carrying a checksum of the best ten levels a side that is checked on each; OKX's from its 400-level book stream (a snapshot, then changes every 100 milliseconds, each naming the one before; perpetual swap sizes are contracts, counted at their face value in coins). A book that misses a change, loses its connection or breaks its sequence is rebuilt from a fresh snapshot, and its seconds out of sync are left out: gaps are never filled. Every second each book is cut into bands by distance from its own mid price, in US dollars (price x size; USDT and USDC count at face value); a period's value is the mean of the book's seconds in it, summed across venues, and a venue counts in a period when its book was sampled for at least half of it (of its part since our collection of the book began; the open hour and day are partials, as far as they went). When our sampler is short of capacity it passes over the books less often than once a second, evenly through the period, and a book is measured against the passes it made. How far a book is complete depends on the coin's price and tick size. Binance's snapshot covers from about 0.15% either side of the price (the busiest perpetuals) to beyond 10% (smaller coins); beyond it the book holds the levels seen to change since it was last synchronised, which are exact, but orders resting unchanged since then are missing, and refresh snapshots every 3 to 30 minutes fill them in. Bybit shows its 1,000 best levels a side and nothing beyond. Kraken shows its 1,000 best spot levels a side and nothing beyond. OKX shows its 400 best levels a side and nothing beyond. Coinbase's stream carries the whole book, kept to 11% either side of the price and complete there. Bands beyond a venue's complete reach are therefore lower bounds, most of all for the largest coins, whose books are densest. History starts when this collection began on 20 October 2026; no earlier values are reconstructed.