Answer · confidence: thin
Which fake-trade detectors were tested on known cheats, and do they match this paper?
Research question: Which published wash-trade detectors have been validated against labeled manipulation cases, and do their volume or transaction-count signals match the complexity indicators in arXiv:2607.13916v1?
The records show that CRASE reports a wash-trading detection F1 of 0.83 on a labelled dataset of 120,000 wallets 16, and it is the only fact that explicitly describes evaluation against labelled cases. Victor and Weintraud's volume-matching algorithm was applied to IDEX and EtherDelta 1 and flags trades where addresses end with the positions they started with 2. The Bitcoin study's volume signal is trade size, with wash traders averaging 5.26 BTC against 7.90 BTC for non-wash trades 5. The verified facts contain no complexity indicators for arXiv:2607.13916v1, so whether these signals match them cannot be determined from the record 516. The newest fact is dated 2025-01-19 16, about 21 months before October 2026.
Checked 2026-10-20. Only documented and supported facts are used.
Dig deeper on this questionSources
[1] documented
Victor and Weintraud (2021) applied a volume-matching algorithm to detect wash trading on decentralized exchanges IDEX and EtherDelta.
"For example, Victor and Weintraud (2021) apply a volume-matching algorithm to decentralized exchanges IDEX and EtherDelta."Direct Evidence of Bitcoin Wash Trading (2024)
[2] documented
The volume-matching algorithm identifies wash trades when multiple addresses execute transactions among themselves and end up with the same market positions they initially had.
"Their method adopts a broader definition of wash trades, so that whenmultipleaddressesexecuteaseriesoftransactionsamongthemselves, afterwhichtheyallend upwiththesamemarketpositionsthattheyinitiallyhad,thenallinvolvedtransactionsaredeemed wash trades."Direct Evidence of Bitcoin Wash Trading (2024)
[3] documented
The study found that 'circle' trades (non-self-self wash trades) accounted for only 1.70% of the total number of self-self wash trades identified in the sample.
"Compared to the 115,275 self-self trades we identify in Section 2, these “circle” trades account for only 1.70% of the total number of (self-self) wash trades"Direct Evidence of Bitcoin Wash Trading (2024)
[4] documented
Adding 'circle' trades to the wash trade sample would increase the share of wash trades within the total number of trades by only 3.53 basis points.
"Given 5,548,867 trades during our sample period, adding “circle” trades to augment the wash trades sample would only increase the share of wash trades within the total number of trades by 3.53 basis points."Direct Evidence of Bitcoin Wash Trading (2024)
[5] documented
Wash traders tend to have smaller average trade sizes compared to regular traders, with an average of 5.26 BTC versus 7.90 BTC for non-wash trades.
"Indeed, the average (median) trade size of non-wash trades is 7.90 (0.62) BTC, while the average (median) trade size of wash trades is 5.26 (0.01) BTC."Direct Evidence of Bitcoin Wash Trading (2024)
[6] documented
Wash trade transactions are particularly prominent in ERC20 cryptocurrencies.
"Wash trade transactions are even more prominent in ERC20 cryptocurrencies."WTEYE: On-chain wash trade detection and quantification for ERC20 cryptocurrencies (2022)
[7] documented
For most ERC20 cryptocurrencies, the rate of wash trade exceeds 15%.
"Our experiments show that for most ERC20 cryptocurrencies, the rate of wash trade reached over 15%."WTEYE: On-chain wash trade detection and quantification for ERC20 cryptocurrencies (2022)
[8] documented
More than 30% of UNI token transactions were identified as wash trades.
"Specifically, over 30% of UNI token transactions were labeled as wash trade."WTEYE: On-chain wash trade detection and quantification for ERC20 cryptocurrencies (2022)
[9] documented
Attackers can amplify the market impact of wash trading by borrowing additional assets from Flash Loan platforms.
"More seriously, attackers can amplify the market impact by borrowing more assets from Flash Loan platforms."Understanding Flash-Loan-based Wash Trading (2022)
[10] supported
Victor et al. conducted the first systematic analysis and detection of wash trading on decentralized exchanges IDEX and EtherDelta.
"Victoretal.[17]conducttheinauguralsystematic analysisanddetectionofwashtradingondecentralizedexchanges IDEXandEtherDelta."Unveiling Wash Trading in Popular NFT Markets (2024)
[11] supported
Cui et al. developed a system for detecting ERC20 token wash trades using algorithms to identify and quantify trading patterns based on on-chain data.
"Basedonon-chain data,Cuietal.[5]deviseasystemfordetectingERC20tokenwash trades,utilizingalgorithmstoidentifyandquantifytradingpatterns."Unveiling Wash Trading in Popular NFT Markets (2024)
[12] supported
Serneels explored wash trade detection from three analytical angles, including closed-loop detection and transaction volume analysis.
"Serneels[15]exploreswashtradedetectionfromthreeanalytical angles, including closed-loop detection and transaction volume analysis."Unveiling Wash Trading in Popular NFT Markets (2024)
[13] documented
Wen et al. developed NFTDisk, a visualization tool for detecting wash trades in the NFT market featuring a Disk Module for address reordering and a Flow Module for multi-level transaction visualization.
"Concurrently,Wenetal[19]developNFTDisk,aninnova- tivevisualizationtoolthatassistsindetectingwashtradesinthe NFTmarket,featuringaDiskModuleforaddressreorderingand visualclutterreductionandaFlowModuleformulti-leveltrans- actionvisualization."Unveiling Wash Trading in Popular NFT Markets (2024)
[14] supported
Son et al. utilized data mining and machine learning to identify fraudulent activities within the NFT market.
"Utilizingdataminingandmachinelearning,Sonetal.[16] delveintotheidentificationoffraudulentactivitieswithintheNFT market."Unveiling Wash Trading in Popular NFT Markets (2024)
[15] supported
Serneels proposed three statistical ways to detect wash trading, using criteria such as closed-loop token trades, closed-loop value trades, and high transaction volumes.
"Serneels introduced three ways to detect wash trading with criterions such as close-loop token trades, closed-loop value trades"Abnormal Trading Detection in the NFT Market (2023)
[16] supported
The CRASE paper reports a wash-trading detection F1 of 0.83 on a labelled dataset of 120,000 wallets, drawn from DeFi exploit post-mortems and influencer fraud reports.
"Evaluation against a labelled dataset of 120,000 wallets drawn from publicly documented DeFi exploit post-mortems and influencer fraud reports yields a sybil detection F1 of 0.88 and a wash-trading detection F1 of 0.83"On-Chain Reputation and Anomaly Scoring: A Cross-EVM Risk Engine for Sybil Detection, Wash-Trading Identification, and Marketplace Trust Operationalisation (2025)
[17] supported
Pelechrinis et al. built a regression model for profit prediction and filtered out transactions that produced above-average profit, as a statistical wash-trade detection approach.
"A more mathematically-based procedure was proposed by Pelechrinis et al. who built a regression model for profit prediction and filtered out anomalous transactions that generate above the average profit [15]."Abnormal Trading Detection in the NFT Market (2023)