How we verify on-chain claims before we publish them

The appeal and the trap

Blockchains are unusual sources. They are public, timestamped and cannot be quietly edited. That makes it tempting to treat anything visible on a block explorer as a verified fact. The ledger does record precisely what moved, when and between which addresses. It records nothing about who controls those addresses or why the transfer happened, and that gap is where most bad on-chain reporting comes from.

So the desk works to a simple rule: the chain confirms the what, and we need separate evidence for the who and the why.

What we check

When a tip or a social media post claims that a fund is selling, an exchange is being drained or a project team has moved its treasury, we go through the same steps.

• Find the transaction ourselves. We do not work from screenshots. We locate the hash on at least two independent explorers or our own node, and confirm status, amounts, token contract and finality.
• Check the token contract. Fake tokens with familiar names are cheap to create. We confirm the contract address against the issuer's own documentation.
• Test the attribution. Address labels on explorers and analytics platforms are useful leads, and they are also sometimes wrong or out of date. We look for a second basis: a published address from the entity itself, a signed message, a court filing, or a consistent history of interactions that supports the label.
• Look at the context. A large transfer to an exchange deposit address may be a sale. It may also be collateral management, an internal wallet reshuffle or a custody migration. We look at what happened before and after, and whether the pattern has occurred before without consequence.
• Ask. We put the specific transaction to the party involved and give them reasonable time to respond. Internal transfers are often explained in one line.

How we word it

If attribution rests on a third-party label, we say so and name the source of the label. If we cannot establish intent, we describe the movement and stop there. We avoid verbs such as dumped when all we can see is a deposit. We include the transaction hash or address in the article so readers can check our work, and if later evidence changes the picture, we update the piece and note the change at the foot.

What we do not publish

Some things fail the test regularly: wallet attributions that trace back to a single anonymous post, claims built on a token with an unverified contract, and alerts about dormant wallets waking up where nothing is known about the owner. These can be interesting. They are not yet news.

The process is slower than reposting an alert. We think a desk that is occasionally second and reliably right is more useful to readers who make financial decisions based on what they read here.