Counting More Efficiently

I’m horrified that this sort of thing is necessary to efficiently query a modern database. (SQL Server in this case.) It feels like something the query planner should be able to do itself.