$2.57 Billion in Fake Volume: Two On-Chain Heuristics You Can Use to Spot Wash Trading Yourself

Since both heuristics are based on behavioral patterns, could they misidentify legitimate high-frequency traders as wash traders?

This is a methodological limitation Chainalysis explicitly acknowledges in its report. The research team specifically noted that MEV bots and arbitrageurs share surface-level similarities with Wash Trading — both involve buying and selling the same Token pair within a very short window. To reduce this kind of misclassification, Heuristic 1 adds the extra condition that "the difference in USD value between the two transactions is under 1%," specifically to exclude genuine Arbitrage — an arbitrageur's profit motive is precisely to capture a price spread, so if the two transaction amounts nearly cancel out, that trade itself captured no spread profit, more consistent with "deliberately manufacturing volume" than "capturing arbitrage profit."

Even so, Chainalysis itself stresses that this methodology reflects "patterns of behavior" rather than "intent," and these heuristics alone aren't sufficient to prove market manipulation actually occurred — they only serve as a starting point for deeper investigation. That's also why, in on-chain analysis, this kind of heuristic is typically labeled "suspected" rather than "confirmed" wash trading, and the same caution should apply when you use them: matching the pattern isn't the same as ironclad proof.

Is the "Token multi-sender" tool mentioned in Heuristic 2 itself a legitimate service? Why does it get exploited for bad purposes?

Yes, token multi-sender tools are entirely legitimate services with genuine use cases — they were originally designed to help users who need to send different tokens to a large number of addresses at once (for example, airdrops or batch payroll payments) do so more efficiently, without manually sending each transfer one by one. There's nothing inherently wrong with the tool itself; the problem lies in how it gets misused.

The reason it gets exploited to evade Wash Trading detection is that this kind of tool inherently possesses the ability to "rapidly disperse funds across a large number of addresses" — which happens to fit the needs of an operator trying to obscure the fact that a single entity is behind everything. Rather than trading repeatedly from one address (which leaves an obvious repetitive pattern), it's more effective to first disperse funds across hundreds of seemingly unrelated addresses, then have each address trade in small volumes independently — on the surface it looks like many independent market participants, but behind it there may be just one operator pulling the strings. This is also why Heuristic 2 needs to additionally trace the original source of funds (the "controller address") rather than only looking at a single address's transaction history — looking at each dispersed address in isolation genuinely makes it hard to spot anything unusual.

When an ordinary person manually checks a Token using a Block Explorer, what specific information should they look at to actually apply the heuristics from this article?

Taking Heuristic 1 as an example, you can search the token's contract address on a Block explorer (like Etherscan), switch to the "Token Transfers" tab, and watch for whether specific addresses repeatedly show up on both the buy and sell side. If the same address shows up as both buyer and seller within a short time window (say, a few minutes), it's worth clicking into that address's page to review its full transaction history — checking whether this pattern repeats multiple times, and whether the amounts on each buy and sell are nearly equal.

Heuristic 2 is harder to fully trace manually, since it requires tracing funds back to their original source, but you can still do a simplified check: if you notice multiple seemingly unrelated addresses that all first appeared around the same time, and whose first incoming transaction all came from the same address, that's the outline of a "multi-sender" pattern worth watching — check whether those addresses subsequently trade the same token. This kind of manual review won't match Chainalysis's scale or precision, but it's enough to build an initial, reasonable basis for suspicion around a token that looks off.

If I discover that a Token I hold shows signs of suspected Wash Trading, should I sell immediately?

It's not advisable to treat "suspected wash trading" as automatically meaning "this token is worthless and should be dumped immediately," but it genuinely should raise your level of scrutiny significantly. The presence of wash trading means at least part of that token's surface-level trading volume is artificially manufactured, which makes it harder to judge how much real market demand actually exists. You might have assumed a token had stable buying support, but if a substantial share of that buying is actually the same set of addresses trading with themselves in a loop, the number of people genuinely willing to buy this token with their own money could be far smaller than the surface figures suggest.

A more practical approach is to treat discovering suspected wash trading as a trigger for reassessment: dig further into whether the token's actual holder count is growing steadily, whether community discussion volume is proportionate to trading volume, and whether the project team has said anything addressing this concern. If further investigation shows wash trading only accounts for a tiny fraction and the token has other solid fundamentals supporting it, panic-selling probably isn't necessary. But if wash trading accounts for a significant share of volume (like the 43% figure in the Volume.li case mentioned in this article), that suggests a substantial portion of this token's apparent market heat may be fabricated, and a more conservative stance is reasonable.