One reason I say that is because they insist on citing Straus' 1988 acyclovir trial (#18), in which 11 patients who received acyclovir reported similar "improvements" to 10 who did not.
Just one problem: The trial was conducted before the 1988 Holmes criteria were developed. In retrospect, it turned out that only six of the original 24 patients recruited actually met the less than stringent Holmes criteria. It's conceivable that only three Holmes patients were among the 21 who finished the trial, and it's unclear how those few Holmes patients (3~6) were randomly distributed between the two arms.