Expert: Chainalysis' poor performance may have put an innocent man behind bars for decades

The recent case of Roman Sterlingov, who was found guilty and sentenced to many years in prison, has raised serious doubts about the reliability of Chainalysis methods. This episode highlights the importance of caution when using such tools in trials, especially when years of imprisonment are at stake.

Bitcoin Fog claims to have processed over 1 million Bitcoins in its lifetime. Through it, hundreds of millions of dollars were allegedly laundered from drug sales from darknet sites such as Silk Road and AlphaBay. Prosecutors said Roman not only used him, but also directed him.

Defense attorneys Thor Ekland and Mike Hassard fought with unrelenting zeal, confronting every argument from the prosecution.

The indictment was largely based on a series of Bitcoin transactions from Sterling's purported Mt. Gox. This series of transactions was subsequently linked to the purchase of the free Bitcoin Fog domain.

But on the other hand, Sterling could have sold bitcoins to someone who purchased the Bitcoin Fog website. It could also be a longer chain of transactions from Sterling to the buyer of the Bitcoin Fog domain.

“Roman Sterling was found guilty on all four counts. I worked on this case pro bono for a year. I will have something to tell, but for now I feel dead.”
March 12 spoke on his Twitter account, Roman’s lawyer, associate professor at the Faculty of Law, crypto consultant and practicing criminologist J.V. Verret:

According to statements by J.V. Verret, the lawsuit alleged that the defendant regularly used Bitcoin Fog and personally ran 2,700 bitcoins through it, while the lawyer pointed to the founder's turnover from 24,000 to 36,000 bitcoins, based on the commissions of the crypto mixer. According to the lawyer, these amounts amount to hundreds of millions of dollars, similar to the case of another head of the cryptocurrency mixer - Larry Harmon and the Helix platform.

But the IRS witness testified that Roman never spent more than $60,000 a year, lived in a one-bedroom apartment and had an income of less than $1.8 million over the past ten years.


Prosecutors K. Alden Pelker and Chris Brown previously wrote a Justice Department publication recommending against basing the prosecution solely on Chainalysis data. They suggested using additional evidence, such as the defendant's possession of the private key to the crypto address under investigation. This is an extremely important aspect, given that Chainalysis heuristics can be wrong in most cases - this is evidenced, for example, by a report from Ciphertrace.

The main problem is that the “co-spend” theory** of Chainalysis assumes that Bitcoins spent together belong to the same user. However, it does not take into account all possible scenarios for using the token. This leads to incorrect conclusions and erroneous assumptions about the relationship between different transactions.

The “peel chain” heuristic assumes that unspent bitcoins are linked into a chain where larger transactions are treated as spent, with change being saved. However, this assumption is not always true, especially if you are sending a large amount of Bitcoin to another person on the chain. It is also impossible to apply this heuristic if you are transferring the private key to another person in an off-chain transaction, which was common in the early years of Bitcoin, which is what is being discussed in this case.

These two tracking heuristics were used by Chainalysis, whose expert claimed that the “secret sauce” in the code of their solution corrected all the shortcomings. But since the code is a trade secret of the company, he could not tell anything in more detail. As a result, there is a risk that an innocent person will be behind bars for many years.