Live crypto liquidations across Binance, Bybit and OKX
SourcesBinance+Bybit+OKXBTC Liquidation Activity | $356k in 1h | 93% Short | No Cascade
In the last hour, 38 liquidation events totaled $356k across Binance, Bybit, and OKX. Short positioning dominated recent activity: shorts accounted for 93% of hourly volume ($330k vs $26k in longs), while over the past 24 hours shorts made up 21% of the $92m total. The 4-hour series shows no liquidation cascade above the $5m threshold during the observation window.
Every crypto liquidation executed on Binance, Bybit and OKX perpetual futures, plotted as it lands. Time on the X axis, price on Y, dot size scales with the USD notional, fill colour marks the side that got force-closed and the outline tags the venue that reported the print. Filter by exchange or minimum size; pause on hover; jump windows from one-minute scalp tape to 24-hour macro view.
How to read the tape
Each dot is one liquidation reported by Binance, Bybit or OKX. The fill marks the side that got force-closed: red when a long was liquidated (price moved down through its level), green when a short went (price moved up). The outline tags the venue — yellow Binance, orange Bybit, cyan OKX — so you can tell whether a cascade is single-exchange or cross-exchange consensus. The size scales logarithmically with USD notional, so a $1 M force-close visibly stands out from a hundred $10 k events without drowning the tape.
Spotting cascades
A cascade reads as a tight cluster of same-colour dots within a few seconds, with rising sizes as each margin layer falls into the next. Cascades mark local extremes more reliably than any funding or OI gauge, because they are the visible footprint of price punching through dense stop-loss layers. Drop into the 1m window to read the rhythm inside the break; jump to 24h to place the cascade against the day's range. A stricter cascade-detection rule lives in the methodology.
Cross-exchange divergence
When Binance prints a cloud of liquidations while Bybit and OKX stay quiet, it is almost always isolated Binance funding that wiped out local longs without touching the same book on the other venues. The inverse — Bybit firing while Binance is silent — usually means a concentrated cluster of leveraged accounts on Bybit. The cross-exchange tape makes that imbalance literal instead of hidden inside an aggregated 24 h number. Cross-check with the funding term structure to confirm the side and magnitude.
Why Hyperliquid liquidations are not on this tape
Hyperliquid is missing from this tape because the venue does not broadcast individual liquidation events over a public WebSocket the way Binance, Bybit and OKX do. Its order book lives on its own L1, so position-level liquidations have to be reconstructed from on-chain events rather than streamed. Until that integration ships, MarketTrace covers Hyperliquid in two adjacent views: depth and walls appear in the footprint chart, and the hourly funding rate sits on Hyperliquid on the funding scoreboard. When Hyperliquid funding diverges sharply at the same time a cascade prints on the CEX tape, that usually marks real cross-venue stress, not an isolated event.
FAQ
What are crypto liquidations?
A crypto liquidation is a forced close: when a leveraged position's margin falls below the exchange's maintenance requirement, the venue closes the position at market instead of waiting for the trader. This tape shows liquidations that already executed on Binance, Bybit and OKX, not predicted or estimated ones.
Longs vs shorts: what does the 24h split mean?
The split is the share of liquidated notional on each side over the last 24 hours. A heavy skew to longs means leveraged buyers were the ones caught on the wrong side of the move. It describes what already happened; it is not a forecast of the next one.
Is this the same as a liquidation heatmap?
No. A liquidation heatmap estimates where liquidations might sit, by modelling leverage clusters that have not been hit yet. This tape shows where liquidations actually happened, print by print, tagged with the venue that reported each one.