Sorry, P2-20 means 19 persons all at 8 units of welfare. The idea was to intuition pump the person affecting restriction: A+ is now strictly better for everyone, including the person who was in A, and so it might be more intuitively costly to say it is in fact A>A+

You may well have thought about all the 'standard' objections to average util in population ethics cases, but just in case not:

Average util seems to me implausible, particularly in different number cases: for example why bringing lives into existence which are positive (even really positive) would be wrong just because they would be below the average of the lives who already exist.

Related to averaging is dealing with separability: if we're just averaging all-person happiness, than whether it is a good thing to bring a person on earth will depend on the wellbeing of aliens in the next super-cluster (if they're happier than us, then anti-natalism seems to follow). Biting the bullet here seems really costly, and I'm not sure what other answers one could give. If you have some in mind, please let me know!