What is the Optimal Density for Trick-or-Treating?

Another Halloween has passed, and this was the first year my son went out trick-or-treating mostly on his own. He's five, so we could let him go free in our apartment building knowing he was safe, and there were lots of other kids doing the same. This is a benefit of density for getting candy--his haul was fantastic and he never had to cross a street (though we went to a neighboring building, too. This was supervised.). However, in high density trick-or-treating neighborhoods the entire event lasts about ten minutes, which is much too short. It really doesn't take long to go floor by floor scoring a small handful of candy at each door. In the span of a couple of hundred feet he hit nine apartments per floor! The kid has way more candy than we want him to, and he expended almost no effort to get it beyond dressing up. I think he even stopped at home in the middle to get a new candy bucket. He declared the event a success, perhaps even the best day of his life. It is certainly the best return on effort he's ever had, and he will be sorely disappointed if he thinks such returns are easy to replicate. The whole thing made me wonder what is the optimal residential density for trick-or-treating. It is perhaps less than Morningside Heights and greater than rural Idaho. I think kids should expend at least one candy bar worth of energy getting their loot.