Aggregated trades (aggTrades)

Definition and fields

When a taker order sweeps several resting orders at the same price, the exchange produces one fill per resting order. Binance's aggregate-trade stream and archives merge those fills into one record: an aggregate id, the price, the summed quantity, the first and last underlying trade ids, a timestamp and a flag stating whether the buyer was the maker. If the buyer was the maker, the seller took liquidity and the trade is sell-aggressive; otherwise it is buy-aggressive.

The representation loses nothing that a footprint needs. Volume per price per side is unchanged by the merge, and delta, computed as ask-side volume minus bid-side volume at each price, comes out identical. What changes is the count: a chart showing the number of events shows aggressive orders, not raw executions.

• One record per taker order per price, quantity summed.
• Maker flag on the buyer gives the aggressor side without a tick rule.
• Ordered ids allow gap detection when a stream drops messages.

Why a footprint prefers this format

• Fewer messages for the same volume, which matters on active pairs during fast sessions.
• Aggressor side is explicit, so delta is a measurement rather than an estimate.
• Historical archives are published in the same format, so rebuilt history matches the live convention.
• Trade-size distributions are per order, closer to what a big-trades filter intends.

In Senzoukria

The Binance Spot and Binance USD-M Perp sources read the aggTrade stream live, and the crypto history path rebuilds bid × ask cells from the exchange's aggTrades archives, an archive that provides both quantity and side for past sessions. On first load the chart imports recent trades, then continues from the stream; when the stream stays empty for a period, the connection reports it instead of drawing flat bars.

Every aggTrade is stored with its notional as well as its base quantity, so the footprint, the Big Trades bubbles and the trade-size indicators can switch units consistently. Because the record is per aggressive order, the trades-count and trade-size indicators on crypto count orders. The crypto guide links to a build log that documents what rebuilding a current day of bid × ask from these archives costs in practice, measured on BTCUSDT Spot.

Common mistakes

• Comparing an aggTrade count with a raw trade count from another feed and reading the gap as missing data.
• Inverting the maker flag: buyer-is-maker means the seller was aggressive.
• Assuming Bybit uses the same aggregation and flag; verify its side field separately.
• Reading a single large aggTrade as one participant's intent; it is one order, and nothing about who sent it is known.