Point of curiosity: the community prediction is, presumably, an arithmetic mean, but I argue that is not a good model for a dataset that almost certainly gets m

Point of curiosity: the community prediction is, presumably, an arithmetic mean, but I argue that is not a good model for a dataset that almost certainly gets more dense closer to the present, creating a gradient out into the future. It would be great to see the geometric mean as well.