Curriculum·G101 Querying On-Chain Data·about 30 min

Aggregating with SQL and Dune

By the end of this lesson you can

→Explain what aggregation does: turn millions of individual events into a pattern by grouping and counting
→Explain how aggregating trades by counterparty exposed the wash trading that inflated LooksRare's volume
→Compute a wash-trading signal by grouping volume by trading pair and comparing to distinct participants
→Choose the grouping that answers your question rather than reading raw totals

Graduate · enrolled learners

This lesson opens with LooksRare wash trading, January 2022.

What happened: LooksRare launched on 10 January 2022 as an NFT marketplace that paid its LOOKS token to users as a share of each day's trading volume. The incentive was gamed almost immediately: traders sold NFTs back and forth between wallets they controlled, paying the 2 percent marketplace fee and gas, purely to generate volume and farm the reward. Reported volume ran into the billions of dollars within weeks, by one count more than 9.5 billion dollars of Ethereum volume against the market leader OpenSea's 4.3 billion for all of January, and taken at face value it looked like an overnight giant. Aggregating the trades told the real story: CryptoSlam flagged any NFT sold and bought back by the same wallet within seven days, found Meebits, Loot and other royalty-free NFTs moving back and forth between the same wallets for upwards of 50 million dollars of ETH each way, and counted more than 8.3 billion dollars of wash trading, about 87 percent of the volume to date by Decrypt's arithmetic. A huge fraction of the reported volume was the same value circling between related addresses rather than genuine economic activity. No individual trade was hidden; the pattern was only visible once the millions of events were grouped and counted.
The decision point: The LooksRare volume was a true number and a false impression at once. Every trade was real and on-chain, so the raw total was correct, and the raw total was also worthless as a measure of genuine demand, because most of it was value circling between the same hands. The gap between the two is closed by aggregation: the analyst does not read the headline total, they group the events by a dimension that answers the question, counterparty, pair, distinct participants, and count. Grouped by counterparty, the wash trading is obvious; read as a headline, it is a record-breaking marketplace. This is the third analyst skill after knowing the chain is evidence and reading its schema: turning millions of individual, honest events into a pattern by choosing the right grouping, because the answer is almost never in the raw total and almost always in how you slice it.

What you will be able to answer

→How was LooksRare volume inflated, and how was it caught?
→What does aggregation do?
→How do you test whether volume reflects real demand?
→Why was LooksRare's total both true and misleading?

Orientation and Year One are open: anyone can read them without an account. From Year Two onward the lessons are for enrolled learners, because progress through the later years only means anything if it is tracked against a record.

It is free. We do not sell the list and there is nothing to buy at the end of it.

Enroll, it is free Sign in Start at the beginning

Sources and review

Confidence high·Volatility medium·Reviewed 2026-09-15·Owner unassigned

Contested

Estimates of the wash-traded share of LooksRare's early volume vary with the definition used. CryptoSlam identified more than 8.3 billion dollars of wash trading, and Decrypt computed that about 87 percent of LooksRare's volume to date matched CryptoSlam's criteria.

Whether LooksRare's volume passed the market leader's depends on the count. Decrypt, citing Dune, gave more than 9.5 billion dollars of Ethereum volume since the 10 January launch against OpenSea's 4.3 billion for January; The Block, the day before, gave LooksRare 1.77 billion for the month so far against OpenSea's roughly 3.11 billion.

Dune's later analysis (hildobby, December 2022), filtering for buyer equals seller, back-and-forth trades, repeat buys of one NFT and wallets first funded by the same address, found wash trades drove 98 percent of LooksRare's volume but only 25 percent of its trades: the same split between a true total and a false impression, measured by grouping.

The lesson uses LooksRare to teach aggregation; the specific groupings named in the worked example (counterparty, pair, distinct participants) are standard analytic slices, not a single published query. CryptoSlam's own test was a resale by the same wallet within seven days, plus a manual review of prices far above the norm.

← Schema literacy: accounts, transactions, logs, and events What a query cannot tell you →