BTC: Spot Sweep-Like Minutes
Shown as its 3 Oct – 1 Nov 2026 sample; the latest readings and full history are included with Pro.
How many minutes of BTC spot trading in each period were sweep-like: a burst of one-sided aggressive volume that ran through a wide price range.
Why use it. Sweeps are how large aggressive orders clear several price levels at once; they often start or end sharp moves. How to read it. Up counts one-sided buying bursts, down one-sided selling bursts. Rising counts mean more urgent, price-moving trading.
Calculation. A minute counts when at least 75% of its volume is on one side, its volume is at least 4 times the hour's median minute and its high-low range at least 3 times the hour's median range. The tape is minute-by-minute here, not order-by-order, so this is a proxy for single sweeping orders. Published for Bitcoin, Ethereum and Chainlink. Source: our own trade-by-trade price-level capture of the public spot trade feeds of Binance, Bybit, Bitget, Gate, KuCoin, HTX, Crypto.com, OKX, Coinbase and Kraken, for BTC's spot markets against USD, USDT, USDC and FDUSD. Every trade is added, in US dollars and on its taker side, to the price level it executed at, on one fixed grid for the coin across all venues: a tick of 1, 2 or 5 times a power of ten, about 0.01% of the price when the collection began; rows of the views are whole numbers of ticks, so every total is an exact sum of trades. Stablecoin quotes (USDT, USDC, FDUSD) count at face value as USD, so venues quoting in different stablecoins share one price axis. Trades missed during a reconnect are not estimated, and venues we do not collect are not included, so volumes understate the whole market; prices on the venues differ slightly (stablecoin quotes, perpetual premiums), so aggregated highs and lows are the extremes of any covered venue. History starts when this collection began on 22 October 2026; no earlier values are reconstructed. OKX, Coinbase and Kraken are included from 23 October 2026, when this collection added them; no earlier values are reconstructed for them.