[Paper Review] A Game of NFTs: Characterizing NFT Wash Trading in the Ethereum Blockchain

Massimo La Morgia, Alessandro Mei+2 more|arXiv (Cornell University)|Dec 2, 2022

Blockchain Technology Applications and Security21 references5 citations

TL;DR

This paper systematically analyzes wash trading in the Ethereum-based NFT market from inception to January 2022, using blockchain data to detect and characterize manipulative trading patterns. It reveals that 5.66% of NFT collections were affected, generating $3.4 billion in artificial volume, with profit primarily driven by exploiting token reward systems—especially on LooksRare—where 84% of volume was wash-traded and 80% of reward-based operations yielded gains, unlike risky NFT flipping.

ABSTRACT

The Non-Fungible Token (NFT) market in the Ethereum blockchain experienced explosive growth in 2021, with a monthly trade volume reaching \$6 billion in January 2022. However, concerns have emerged about possible wash trading, a form of market manipulation in which one party repeatedly trades an NFT to inflate its volume artificially. Our research examines the effects of wash trading on the NFT market in Ethereum from the beginning until January 2022, using multiple approaches. We find that wash trading affects 5.66% of all NFT collections, with a total artificial volume of \$3,406,110,774. We look at two ways to profit from wash trading: Artificially increasing the price of the NFT and taking advantage of the token reward systems provided by some marketplaces. Our findings show that exploiting the token reward systems of NFTMs is much more profitable (mean gain of successful operations is \$1.055M on LooksRare), more likely to succeed (more than 80% of operations), and less risky than reselling an NFT at a higher price using wash trading (50% of activities result in a loss). Our research highlights that wash trading is frequent in Ethereum and that NFTMs should implement protective mechanisms to stop such illicit behavior.

Motivation & Objective

• To systematically detect and characterize wash trading in the Ethereum NFT market across all ERC-721 NFTs from inception to January 2022.
• To analyze the economic incentives behind wash trading, particularly the role of marketplace token reward systems in driving manipulation.
• To evaluate the profitability and risk profile of different wash trading strategies: inflating NFT prices versus exploiting reward mechanisms.
• To identify patterns in wash trading behavior, including duration, frequency, and actor coordination.
• To provide actionable insights for NFT marketplaces to implement protective mechanisms against such manipulative practices.

Proposed method

• Collected and parsed all ERC-721 NFT transfers on the Ethereum blockchain from inception to January 18, 2022, covering 34.75 million assets and $34B in transaction volume.
• Applied and compared multiple wash trading detection techniques, including round-trip trading analysis, common funding and profit accounts, and graph-based methods on transfer, payment, and sales networks.
• Identified wash trading events by detecting two-party round-trip trades with shared funding and profit accounts, confirming collusion through transaction clustering and timing patterns.
• Quantified artificial volume by isolating transactions attributable to wash trading operations across six major NFT marketplaces.
• Classified wash trading activities into two categories: (1) price manipulation via reselling at inflated prices, and (2) exploitation of volume-based token reward systems.
• Used statistical analysis to evaluate profitability, loss rates, and success probabilities for each strategy across detected operations.

Experimental results

Research questions

• RQ1 What is the scale and temporal pattern of wash trading across the Ethereum NFT ecosystem from 2015 to early 2022?
• RQ2 How do wash trading operations differ in structure and coordination, particularly in terms of actor behavior and transaction timing?
• RQ3 To what extent do marketplace token reward systems incentivize and enable wash trading, and how do they compare economically to price manipulation strategies?
• RQ4 What is the actual profitability and risk profile of wash trading when targeting NFT price inflation versus reward system exploitation?
• RQ5 Which NFT marketplaces are most vulnerable to wash trading, and what structural features make them susceptible?

Key findings

• Wash trading affected 5.66% of all NFT collections, generating $3.406 billion in artificial trading volume across the Ethereum blockchain.
• More than 80% of wash trading operations targeting token reward systems were profitable, with a mean gain of $1.055 million on LooksRare, compared to only 50% success rate for price-inflation strategies.
• Over 84% of LooksRare’s total trading volume was attributed to wash trading, making it the most affected marketplace, primarily due to its volume-based token reward mechanism.
• 59.86% of wash trading events followed a round-trip pattern between two accounts, with 25.98% lasting only one day and 51.67% lasting less than ten days.
• 27.16% of wash trading accounts were responsible for 72.93% of all detected activities, indicating the presence of serial wash traders with coordinated behavior.
• The study confirms that exploiting reward systems is significantly less risky and more profitable than reselling NFTs after artificial price inflation, which resulted in losses in half of the attempts.