The SPM took it on the chin in Week 3, going 17-24 (41.46%) on all games (excluding those involving FCS teams and those for which there were insufficient comps). Above what is usually a magic confidence level, it went 6-8 (42.86%) and within the usual magic confidence range, it managed only 3-3 (50%).
Last week, we pulled out all of the games involving FCS opponents so that we could monitor them separately. The early returns are mixed: It was better over the confidence level and in the confidence range as you’d expect, but slightly worse overall. It was a weird week, though, so we’ll just keep monitoring. Also, this was the first week we actually compiled results separately depending on whether the SPM complained about not having at least two good comps. The official results we post here are those that don’t include the games involving FCS opponents or those for which the SPM doesn’t have at least two good comps. That only starts in Week 3, though, as the official results for Weeks 1 and 2 include both of those categories as of now.
For the season so far, the SPM is 61-69 (46.92%) overall, 33-32 (50.77%) over the confidence threshold, and 17-11 (60.71%) within the confidence range.
Although it was a game involving an FCS opponent and thus didn’t count toward the official results, the SPM did get the Tennessee-Chattanooga game right, saying the Vols would easily cover the 24-28 point spread. Let’s take a look to see what it says this week about the Gators
Vols-Gators
We’re only three weeks into the 2019 season, and although there are now enough comps to work with, some of them are still a bit questionable. For that reason, we’re still using a combination of 2018 and 2019 data and weighing them accordingly.
2018
Without going through all of it, the 2018 data spits out an estimated score of Florida 41.4, Tennessee 18.9. That uses spot-on comps, but it’s last year’s teams.
From the perspective of 2019 Tennessee
Tennessee scoring offense for the season: 33.7
Florida scoring defense for the season: 13.7
The Florida scoring defense is most similar to the following prior Tennessee opponent(s) (FBS only):
BYU 27.7
Georgia State 43
Against BYU, Tennessee scored 26 points.
Against Georgia State, Tennessee scored 30 points.
Estimated points for Tennessee against Florida: 28
Those are bad scoring defense comps, with BYU and Georgia State both being significantly worse than Florida, so assuming the Vols are going to score about what they scored against those teams is suspect. Bottom line: Be wary of that 28 points for Tennessee.
Tennessee scoring defense for the season: 22.3
Florida scoring offense for the season: 32.7
The Florida scoring offense is most similar to the following prior Tennessee opponent(s):
Georgia State 32
BYU 23.7
Against BYU, Tennessee allowed 29 points.
Against Georgia State, Tennessee allowed 38 points.
Estimated points for Florida against Tennessee: 33.5
Estimated score: Tennessee 28, Florida 33.5
From the perspective of 2019 Florida
Florida scoring offense for the season: 32.7
Tennessee scoring defense for the season: 22.3
The Tennessee scoring defense is most similar to the following prior Florida opponent(s) (FBS only):
Kentucky 23.3
Miami (Florida) 17.3
Against Kentucky, Florida scored 29 points.
Against Miami (Florida), Florida scored 24 points.
Estimated points for Florida against Tennessee: 26.5
Florida scoring defense for the season: 13.7
Tennessee scoring offense for the season: 33.7
The Tennessee scoring offense is most similar to the following prior Florida opponent(s):
Kentucky 32.3
Miami (Florida) 36
Against Kentucky, Florida allowed 21 points.
Against Miami (Florida), Florida allowed 20 points.
Estimated points for Tennessee against Florida: 20.5
That’s much better data to use to determine Tennessee points.
Estimated score: Florida 26.5, Tennessee 20.5
SPM Final Estimates
Putting all of that together, here’s what the SPM gets:
SPM Final estimated score: Tennessee 23.1, Florida 33.8
SPM Final estimated spread: Florida -10.7
SPM Confidence level: 3.3
That confidence level puts it under our threshold for feeling very good about it, at least against the spread.
Eyeball adjustments
That looks about right for Florida’s points, but seems a wee bit high on Tennessee’s points . For that reason, my eyeball-adjusted prediction is Florida 34, Tennessee 20.
Other predictions from other systems
The Vegas line has Florida as the favorite at between -14 and -14.5 with an over/under of 48.5 – 49. That translates to something approximating Tennessee 17, Florida 31.
Bill Connelly’s SP+ likes Florida 35-17 and gives the Vols a 14% chance of winning. SP+ is 54.3% overall and 64% on games for which it and the spread are at least three points apart.
As I said above, for the season, our SPM is 61-69 (46.92%) overall, 33-32 (50.77%) over the confidence threshold, and 17-11 (60.71%) within the confidence range. That would put our overall number 38th on this projection tracker. Our favorites-only number would be 1st. Well, second behind Bill’s and only until everyone else was allowed to use their favorites-only number.
ESPN’s FPI gives the Vols a 13.9% chance of winning.
Bottom line
Florida by 14 seems about right. Let’s hope right is wrong this week.
What are y’all thinking?