That's essentially average-utilitarian (has the happiness of a random person in the population gone up?); I guess many people are closer to total-utilitarian (ie. the total number of happy/surviving people is also important).
Wouldn't total-utilitiarian then prefer a population of 1000000 mostly unhappy people over a population of 100 very happy people? In that case it would at least be an awkward benchmark, as you just need to procreate as much as possible to get the "best" scenario.
Yes, there's an interesting thing on that here:
http://plato.stanford.edu/entries/repugnant-conclusion/
I haven't looked at enough philosophy to be able to figure out what I'd place myself as, average-utilarianism also seems to lead to morally-counter-intuitive stuff.