we could probably test...
get a series of Icke publications, picked at random from a wider pool of (hopefully) all of his work, find all the verifiable claims made in each, and tally up the percentages.
I'd be surprised to find anyone who stands up to a 90% truth standard after such though