Opposing counsel produces a report. It has transaction graphs, exchange names, wallet addresses, and a bottom-line number: the other side is concealing $420,000 in Bitcoin. The analysis cites the blockchain, which everyone knows is permanent and tamper-proof. The figure starts to feel less like an opinion and more like a measurement.
It is not. A blockchain tracing report is not a bank statement. It is expert opinion, and the ledger underneath it is the only part that qualifies as fact. Almost everything that turns raw ledger entries into a conclusion about who owns what is interpretation, and interpretation is exactly what the rules of evidence let you test.
A tracing report is expert opinion, and the standard just tightened
Because a tracing report is offered as expert testimony, it has to clear the reliability bar that governs all expert testimony. In federal court and in the large majority of states, that bar is Daubert, which asks whether the method can be and has been tested, whether it has a known error rate, whether it has been subjected to peer review, and whether it is generally accepted in its field. A minority of states still apply the older Frye standard of general acceptance. Under either test, the report is challengeable on its method, not merely on its arithmetic.
This matters more now than it did two years ago. Effective December 1, 2023, Federal Rule of Evidence 702 was amended for the specific purpose of curbing expert overstatement. The proponent of the testimony now must show, by a preponderance of the evidence, that the opinion reflects a reliable application of the method to the facts of the case. The advisory committee named the target plainly: experts overstating the certainty of their conclusions beyond what the methodology, properly applied, can support. A confident cryptocurrency tracing report is very often an exercise in precisely that kind of overstatement. What follows is a field guide to where it tends to happen.
1. Clustering is a heuristic, not a fact
The first thing a tracing tool does is group many individual addresses into a single wallet attributed to one party. That grouping is not read off the chain. It is inferred, primarily through the common-input-ownership heuristic: when a transaction spends several inputs at once, the tool assumes one party controlled all of them, because ordinarily you would need every private key to sign. That assumption is usually correct. It is not always correct.
Collaborative transaction types such as CoinJoin deliberately combine inputs from unrelated people into one transaction, and some exchanges batch many customers' withdrawals the same way. When that happens, the heuristic can fold addresses into a cluster the named party never controlled. Tools layer a second set of change-address heuristics on top to guess which output is the sender's change rather than a payment, and those are probabilistic too. Now connect this to Daubert: one of its reliability factors is a known error rate, and commercial blockchain analytics tools generally do not publish one for their clustering. A method whose error rate is unknown, applied to reach a confident conclusion, is the overstatement the amended Rule 702 was written to catch. Ask whether any cluster central to the report depends on these assumptions, and whether commingled or collaborative transactions could have contaminated it.
2. The attribution labels come from a proprietary database
When the report states that funds went to a named exchange, or were cashed out at a particular service, that label did not come from the blockchain either. It came from the vendor's proprietary attribution database, assembled through the vendor's own private data collection and never made public. Opposing counsel cannot audit how a given label was assigned, when it was last verified, or what confidence the vendor itself attaches to it.
This is a sufficiency-of-the-basis problem. The expert is resting an opinion on a third-party dataset they did not build and cannot fully validate, which is the kind of foundation Rule 702 asks the proponent to justify. A label can also be stale, or it can be correct about the exchange while telling you nothing about which customer account actually received the funds. Ask for the provenance of every entity attribution the conclusion depends on, and ask the expert to separate, clearly, what they personally verified from what they accepted from the tool.
3. An address is not control
This is the single point that undoes the most tracing reports, and it is simple. On-chain data can show that funds moved to an address. It cannot show who holds the private key to that address.
An address is not control.
If the report concludes that a party still holds a quantity of Bitcoin because the coins sit at a particular address, that conclusion quietly assumes key control the chain cannot establish. Funds resting at an exchange deposit address are controlled by the exchange, not the depositor, and they exist as an account balance the exchange custodies. Coins can be sent to an address the sender does not control at all. Before traced to becomes owned by, someone has to prove custody of the keys, and that proof lives off the blockchain, in account records, devices, and testimony.
4. The tracing method chooses the answer
Once traced funds flow into an address that already holds other money, the coins are commingled, and there is no longer a single factual answer to which dollars went where next. There are only methods. First-in-first-out, last-in-first-out, and proportional or haircut tracing each follow value through a commingled pool differently, and they routinely produce different conclusions about which downstream output carries the traced funds.
A report rarely tells you which method it used, yet the choice can move the bottom-line number materially. This is a reliable-application question under Rule 702(d): if switching from one accepted method to another, equally accepted one erases the conclusion, the method was driving the result, not the evidence. Ask which tracing methodology the expert applied through any commingled address, and ask whether a different standard method would change the answer.
5. What the report cannot see
Blockchain analysis only sees the chains the tool indexes and the activity those chains record. Several things sit outside that view entirely. Privacy-focused assets such as Monero are not traceable with these tools. Chains the vendor does not cover are invisible. Cross-chain bridges can obscure where value re-emerges. And transfers between two accounts inside the same exchange settle on the exchange's internal ledger and never touch a public chain at all.
A report that claims the cryptocurrency is fully accounted for is making a statement it cannot support, because it cannot see what left the visible surface. The honest version of that claim is narrower: here is what the indexed chains show. Watch for completeness language that outruns the method, and hold it to the standard the rule now demands.
6. The dollar figure is a set of assumptions
Finally, the headline valuation is not a reading off the chain. It depends on a chosen valuation date and a chosen price source. Cryptocurrency is volatile enough that the date used can swing the number significantly, and different price indices disagree at any given moment.
The date is not a free choice, and it varies by jurisdiction. Most states divide marital property under an equitable-distribution standard, while nine community-property states characterize and value marital assets differently, and the controlling valuation date, whether date of filing, date of separation, or date of trial, differs from state to state. Confirm which date and which price source the expert used, and whether those choices match the valuation standard your jurisdiction actually requires.
This is not only a divorce problem
The same scrutiny applies anywhere an opposing cryptocurrency tracing report shows up. It applies in civil fraud and asset-recovery litigation, where a tracing report supports a claim about where money went. It applies in probate, where an estate's digital assets have to be identified and valued. It applies in bankruptcy, where concealed-asset disputes increasingly involve crypto. The report is expert opinion in all of those settings, and the same reliability standards govern it.
Why this is landing on more desks
Cryptocurrency used to be a niche issue. It is becoming a routine one. Federal reporting that took effect in 2025 now requires digital-asset brokers to issue 1099-DA forms, which means undisclosed holdings increasingly leave a paper trail that surfaces in discovery, and court rules around the country are catching up to the asset class. The practical result is more tracing reports filed in more matters. Most attorneys have never had anyone qualified to tell them whether the report in front of them is sound.
What this adds up to
The blockchain is evidence. The report written about it is argument. The ledger entries are real, but the clustering, the attribution, the assumption of key control, the tracing method, the completeness claims, and the valuation are all interpretive layers stacked on top, and each is a place where a confident report can be wrong. Since December 2023, the rules of evidence have explicitly told courts not to let an expert's confidence outrun the method. A tracing report deserves the same scrutiny you would give any other expert opinion, not deference because it invokes the word immutable.
I review cryptocurrency tracing reports for exactly these failure points, as a consulting expert first and a testifying expert where the matter warrants it. If you have received a blockchain analysis you need vetted, or you suspect an opposing expert has overstated what the chain can prove, I am available to look at it.