Back in May, Bill Connelly updated his pre-season SP+ projections and Baylor saw its ranking fall another ten spots down to #40. This surprised many Baylor fans, especially those who hold Big 12 Championship and CFB playoff aspirations. How can the Sugar Bowl winner, reigning Big 12 Champion, and pre-season favorite to win the conference have such a poor outlook according to college football’s most popular statistician?
If you read my Primer on Predictive Statistics, you know that returning production, recent recruiting success, and historical performance are the three components in SP+’s pre-season calculations. We reached out to Bill with questions about how Baylor fares in each of these categories. Here’s what he told us:
ODB: Returning production is the biggest piece of your pre-season ratings. For a team like Baylor, who lost 77% of their passing yards when Gerry Bohanon announced he was transferring, that obviously hurts a lot. However, Bohanon was beat out in camp by Blake Shapen, suggesting a potential upgrade at the position. Does your model treat this type of lost production differently than, say, a QB leaving for the NFL draft? Do you think it matters?
Bill: There’s no differentiation regarding why a player leaves. It’s simpler that way, and the Shapen-Bohanon types of cases are extremely rare. I’m at peace with it — the simpler, the better — but obviously it gives a potentially misleading impression of the QB situation at BU this year. (It also reinforces how unique a decision Aranda made and how much pressure there is on Shapen to thrive.)
ODB: Do recruiting ratings interact with returning production in your pre-season ratings? For example, Baylor lost a lot of production at WR, but they picked up two WRs in the last class, including Armani Wienfield, their highest rated recruit. What about incoming transfers’ positions and returning production?
Bill: Again, the simpler, the better. I’ve always had grand designs of coming up with weighted recruiting class rankings — weighted toward the positions that are most likely to make a quick impact — but haven’t yet. So no, the projections just know what Baylor lost and know the average talent Baylor signed.
ODB: What goes into the “recent history” component of pre-season ratings? Is it just the season-ending overall SP+ rating, or do some historic factors get reweighted based on their correlation over time? For example, Baylor did very well generating turnovers last year, but turnovers are less correlated over time than efficiency. Does the recent turnover rate get downweighted to compensate?
Bill: Recent history just means a weighted average of the last five years of SP+ ratings. Last year carries the most weight, the year before carries a decent amount and five years ago carries very little. So for your reference to turnovers, Baylor didn’t get a ton of credit for them — based on national averages for fumble recovery rates and the ratio of INTs to PBUs, their TO margin should’ve probably been closer to +5 than +12 — and it was likely one of the reasons they ended up in the teens instead of the top 10 last year. But that teens ranking is the biggest piece of their recent history average.
A big thank you to Bill Connelly for answering our questions. Follow him on Twitter at @ESPN_BillC and stay tuned to OurDailyBears.com for more Baylor football content!