Are you sure about that? I thought it does prevent infection with wild-type virus. In general herpesviruses don’t superinfect well (though it does happen) and since the vaccine virus establishes its own persistent infection I’d expect that superinfection with wild-type virus would be rare. If I’m interpreting a number of papers correctly, this is correct — superinfection with wild-type virus does occur but it’s unusual, a few percent, and associated with reduced disease — but almost everything I find does look more at disease than evidence of infection.

Testing for superinfection would probably be very difficult, because you can’t use the usual seroconversion measures; you’d have to actually sequence out virus and confirm that it’s wild-type and not vaccine strain. I believe that it’s this technical difficulty that’s led to the FDA definition of preventing disease rather than infection.