Study Reveals Widespread Wash Trading in NFT Markets

The latest academic analysis confirms that a large share of activity on major NFT exchanges is likely artificial, with machine‑learning models flagging roughly 38% of trades and 60% of total value as wash trading [2]. Earlier blockchain‑focused work had already highlighted that, at its peak in January 2022, more than 80% of reported volume on Ethereum‑based NFT platforms was wash‑traded, leaving a persistent 58% for the full year [1].

Key takeaways

• Around 38% of NFT transactions show patterns consistent with wash trading [2].
• These suspect trades account for about 60% of the total monetary value exchanged on major NFT marketplaces [2].
• Prior on‑chain analysis found wash‑traded volume peaked at over 80% in January 2022 and remained above half of annual volume [1].
• Researchers used filters such as self‑trades, back‑and‑forth swaps, circular loops, and common funding to train detection models [2].
• The prevalence of wash trading undermines the reliability of trade‑volume metrics that many platforms and investors rely on [1].

Machine‑Learning Detection of Wash Trades

The Boston University study, led by Gerry Tsoukalas and collaborators from the University of Pennsylvania and UCLA, built a set of on‑chain detection filters to flag suspicious behavior. These filters targeted self‑trades (where the same user controls both buyer and seller), back‑and‑forth swaps, circular trade loops involving three or more wallets, and cases where both parties were funded by the same source [2]. The filtered transactions served as training data for machine‑learning models that could identify broader patterns of manipulation across the blockchain.

Echoes of Earlier On‑Chain Analyses

The findings align with earlier work that manually examined Ethereum NFT activity. That analysis reported that at its height in January 2022, more than 80% of NFT trade volume was wash‑traded, and the practice persisted at a 58% rate for the entire year [1]. Both studies note that token‑reward programs on platforms such as LooksRare, X2Y2, and Blur incentivize users to generate artificial volume, often by moving the same NFT between wallets they control. While some participants profit from these schemes, many incur net losses due to gas fees, yet the inflated volume skews market metrics [1].

Why it matters

Wash trading erodes confidence in NFT market data, making trade volume an unreliable proxy for genuine demand. As analysts and investors increasingly depend on on‑chain metrics to assess platform health, inflated figures can mislead decisions and obscure the true economic activity behind digital collectibles. The new machine‑learning approach offers a more systematic way to filter out inorganic trades, but the study also underscores the need for broader transparency and possibly regulatory oversight as the NFT ecosystem matures. Continued refinement of detection methods and greater awareness of manipulation tactics will be essential for building a trustworthy NFT market.