this is another interesting comment on the issue: https://news.ycombinator.com/item?id=41977430

but crucially, it assumes that variations in submissions and voting activity are strongly correlated, which I imagine is not necessarily the case