A season-to-date pace average treats October like gospel. A pace prior — shrunk toward a roster-informed expectation — cuts our totals error by about two points a game in the season’s first six weeks.
The distinction sounds academic. Two points a game is the difference between a model that pays and one that bleeds.
