I think I accounted for the baby-boomer age-affect by comparing the death-rate to the linear-model’s prediction. ie I don’t see how those dying in 2013, 2014, 2015 could fit a linear trend, but then 2016 buck that trend except by a fluke. Death-rate suddenly turning exponential in 2016 doesn’t sound plausible to me.

I think you’re saying that I could have considered whether 2016’s celebrities died especially young given their ages? That’s an interesting question, and I don’t have an answer. Whether they did or not doesn’t really answer whether unusually many died or not, though.

Sorry if I misinterpreted your comment though!