Thanks, but can you clarify... does that mean probing (a..z)×(a..z) (one pair per probe), so there's at least a giant increase in probing required per character

Thanks, but can you clarify... does that mean probing (a..z)×(a..z) (one pair per probe), so there's at least a giant increase in probing required per character? And perhaps even more each character in, since probing for the Nth character now requires (a..z)^(N-1) × (a..z)×(a..z) ? (I'm guessing also, though, it may be possible to probabilistically probe multiple ranges of the secret at once... in a process that seems vaguely similar to forward-error-correction coding.)