[ok_craig]

Can you explain why the signature matching one in the blockchain disproves the theory?

I imagine the signature must also be somehow related to his private/public key pair otherwise his blog is just junk cryptography and would have been easily debunked anyway?

[murbul]

Signatures are unique to the message that's being signed, otherwise they'd be useless.

[amdavidson]

Signatures are only probably unique. Theoretically there can be collisions.

Granted, the likelihood that he inadvertently collided with a public transaction is astronomically small.