Thanks again, that makes perfect sense.
It is a little annoying that we have very little chance of finding primes on those WUs though.
If it was found to be prime by the initial finder we would miss out and only if the initial test had incorrectly shown it as non-prime and we found it to be a prime would we get an elusive prime against our name. I suspect that we have never found a prime this way. Therefore I'd rather not be spending my time on WUs that have an extremely low chance of being a prime - but we have committed to do this work so I have no choice (other than to move to a different sub-project I guess?)
And what about hosts like mine that are, at the moment, subject to AR. The WUs I get as initial tester currently show a minimum quorum of 1 so they aren't being double checked by anyone? Is the statement on Ray and Mark's site just plain wrong?