Short answer
Whale flows are the net value of unusually large trades above a per-coin size threshold: large taker buys minus large taker sells. They show where big orders hit the market, but a single print can be a hedge, an arbitrage leg or an exchange rebalance, so large-trade flow is context, not a signal on its own.
Last checked
Key facts
| What is counted | Prints above a per-coin size threshold |
| Net flow | Large taker buys − large taker sells (quote-currency notional, USDT on Binance USDⓈ-M, roughly USD) |
| Coverage | Partial view: only the trades the monitor reads |
| Common misread | Hedges and arbitrage look like directional bets |
| Better with | Funding, OI and price reaction |
How does a large-trade monitor work?
A typical monitor listens to trade prints and keeps only those above a size threshold. Every trade has a buyer and a seller, so a monitor gives each print a direction by its taker, the side that crossed the spread: a taker buy lifted the ask, a taker sell hit the bid. Many exchange trade feeds mark each print with a flag showing whether the buyer was the maker. If it was, the taker was the seller. Check your exchange's API documentation for the exact field name.
The threshold is often set per coin, because 1 million dollars is large for a small coin and ordinary for BTC. A threshold applies to each print, not to each order. How fills are grouped into prints depends on the feed the monitor reads. A big market order that sweeps several price levels can show up as several smaller prints. If each is below the threshold, the monitor counts none of them.
Net flow = sum of taker-buy notional above the threshold − sum of taker-sell notional above the threshold, over a fixed window. Notional = price × quantity, measured in the quote currency. On Binance USDⓈ-M contracts that is USDT, which is roughly dollars. Monitors differ on whether a print exactly at the threshold counts. In this article it does not, so a print of exactly 1 million dollars is ignored.
Example: take a 1 million dollar threshold and a BTC price of 80,726.7 USDT. The cutoff quantity is 1,000,000 divided by 80,726.7. A print slightly below that quantity has a notional just under 1 million, so it is ignored. A print slightly above it has a notional just over 1 million, so it counts.
Now suppose the window holds one more qualifying taker buy than qualifying taker sell, and every print has the same size, just above the cutoff, at that price. The buys and sells cancel in pairs. Net flow is positive and equals the notional of a single print, a little over 1 million dollars. You can run the same sum on your own data to check a monitor's figure.
1 million dollars is used here only to keep the arithmetic simple. Because that size is ordinary for BTC, a real BTC threshold would usually be set higher.
How do thresholds change the picture?
The threshold decides what counts as a whale. Set it too low and ordinary trading noise floods the data. Set it too high and only a handful of prints remain, which makes the net figure swing wildly. A per-coin threshold scaled to typical trade size keeps coins comparable. When you read anyone's whale data, including ours, ask what threshold and time window were used. The same market can show opposite net flow over one hour and over twenty-four hours.
Why can whale flow mislead?
Big players hedge, rebalance and arbitrage across venues. A large sell on one exchange can be matched by a buy elsewhere. Flow on one venue is therefore a partial view.
One check: see whether a matching buy shows up on another venue or in spot around the same time, and whether price followed. A large sell with no follow-through is not proof of a directional bet.
How should I use it?
Ask whether price reacted. Heavy buying that fails to lift price suggests sellers absorbed it. Selling that does not push price down suggests the same for buyers.
Combine flow with funding and open interest to see if leverage is joining the move. If net buying comes with rising open interest and funding turning more positive, that is consistent with new leveraged longs joining, though open interest alone does not tell you which side opened. If open interest falls while flow is large, positions may be closing rather than opening.
Flow does not tell you where price goes next, and acting on it with leverage can lose your full margin.
What does our dashboard show?
Our daily dashboard publishes net large-trade flow per coin from our own monitor. It covers only the trades our monitor reads, so it is a partial view and does not show the whole market. Check the dashboard page for the window, the per-coin thresholds and whether a print exactly at the threshold counts there.
We do not estimate liquidations because no complete public liquidation feed exists. Binance's liquidation stream pushes only the latest liquidation order per symbol within each 1000ms window, so some liquidations never appear.
FAQ
Do whales move the market?
Large orders can move thin markets briefly. Sustained moves need broad participation.
Is on-chain whale tracking the same?
No. On-chain transfers show wallet movements, not exchange trades. Exchange matching happens off-chain.
Can I follow whale trades profitably?
There is no evidence that copying single large prints works reliably. Treat flow as context, not a trade signal. With leverage, a small adverse move can wipe out your margin and trigger liquidation.
Why do whale flows differ between exchanges?
Each venue sees only its own trades. Large players often split orders across several venues, so one exchange can show net buying while another shows net selling at the same time.
Sources
Figures change. Check the primary sources below.