This is the 5th year I’ve done this RB strength of schedule projection … and it’s time to blow up the process.
We’re always looking for improvement and refinement, of course, here and in anything else we do. That often means adding data. In this case, though, I think we can actually do more with less.
It’s tough to find sticky year-to-year stats for what defenses are allowing to RBs. What does “sticky” mean? That the results carry over year to year, so that a defense’s performance 1 year might help us predict what it will do the following season.
I previously found that rushing attempts, rushing yards and PPR points were most correlated year to year. But even those links weren’t all that strong. And looking at a defense’s performance against the run (primarily via Football Outsiders’ DVOA) has offered negligible value for projecting.
What I finally learned this time around is that I was probably trying too hard. Rather than tracking all these categories, it’s probably better to just look at game script.
The year-to-year correlations on the 3 aforementioned categories have fallen in the 0.34 to 0.37 range, on average, dating back to 2009. With 0 signaling no correlation and 1 (or minus-1) signaling a direct correlation, that’s obviously an area of limited strength.
This time around, though, I also compared RB rushing yards allowed to team wins. Why yards? Because they’re closely tied to carry volume and the most predictable driver of fantasy-point scoring at the position. (TDs are obviously big in season but less predictable year to year.)
From last year back through 2010, the median correlation of yards allowed to team wins (by the defensive team) is -0.5. That inverse correlation means the worse a team is, the more rushing yards it’s likely to allow.