Why numeric projections may undersell Florida football again

Shortly before last season ended, I noted that the major data-driven projection systems were seeing Florida winning about seven or eight games. It was close enough to the season that optimism was high among Gators, though, and in GC’s Bullgator Den, a poll showed members predicting more like eight or nine wins. I gave reasons why each could’ve been correct. In actuality, UF ended up winning nine regular season games, showing that the fans were closer than the computers last year.

This year the S&P+ system predicts Florida winning 8.3 games, while ESPN’s FPI sees 8.7 wins. The FEI system hasn’t put out its projection yet, but last year it came in between those two. The consensus is around eight or nine wins.

A tougher schedule that trades Colorado State for Miami and what turned out to be a one-dimensional Mississippi State team for a likely more complete Auburn is one reason why UF might fail to hit the nine win threshold again. Eight or nine wins just might be the right expectation for 2019.

I do think the Gators have a chance to outperform that expectation, and here are the reasons why.

Team Talent

One factor that prediction systems use is the amount of talent on the team. Often it’ll come in the form of past recruiting class rankings, but the 247 Sports Composite Team Talent Ranking does just fine as well. It hasn’t been updated for 2019 yet, but the 2018 one can still tell us some things.

One thing that probably won’t change from last year to this year is FSU being ranked ahead of Florida. Yet as we saw last year, that program is not living up to its talent ratings for a variety of reasons. The Seminoles should be better this year, but they won’t be top ten good. That opponent is being oversold by talent ratings right now.

The Gators were a little ahead of but in the same neighborhood talent-wise as Tennessee, Auburn, and Miami last year. The Vols and Hurricanes have new offensive systems, though, and UF’s starters have a higher average ranking than the Tigers’ do.

That said, if a system does just use recent class rankings, then UF might be getting a little more credit than it should. The 2017 class has lost three of its 4-star players (James Robinson, Daquon Green, Kadeem Telfort), the 2018 class has lost two (Justin Watkins, Malik Langham), and the 2019 class has already lost four (Chris Steele, Jalon Jones, Diwun Black, Deyavie Hammond).

I did mention last year that the Gators had a number of 3-star guys who played better than that rating implies. They still have a few like Lamical Perine, David Reese II, and Jabari Zuniga, but there are fewer of them with Jachai Polite, Jawaan Taylor, and Vosean Joseph in the NFL.

This factor is probably a wash, with UF ending up around where it should be.

Returning Production

The next thing these systems account for is what returns from the prior year. Ideally a system will look at returning production like S&P+ does and not just returning starters or number of starts.

Last year, Jordan Scarlett, Van Jefferson, and Trevon Grimes weren’t counted in returning production stats because they didn’t play for Florida in 2017. This year, returning production won’t count the impact of a fully healthy Malik Davis, a full season (ideally) for Marco Wilson, or the expected excellence of Jonathan Greenard on defense.

Wilson is a big upgrade over the combination of true freshman Trey Dean and CJ McWilliams at corner. Davis was elusive and explosive running behind a bad offensive line in 2017, and while he looked tentative coming off of his knee injury in 2018, he should be fully ready to go this year. If Greenard can mostly return to his 2017 form when he had 15.5 tackles for loss, he will fill in much better for Polite than the generic replacement player that these systems very well could be assuming.

These returning production measures also will not count anything for Brett Heggie starting some games in 2017. They will see four new OL starters who didn’t start a single game or play a significant number of snaps last year. They will not see a more accurate picture, which is three new starters plus one with an asterisk for having significant experience from two years ago.

Those are only four players among many, but especially the two defenders are likely to make outsized impacts.

Recent Performance

One of the strongest year-to-year correlations in college football is actually straight up winning percentage. One of the best predictors of how team will do this year is what that team did last year. That’s why if you predict all conference champions and playoff teams to be the same in a given year as in the preceding one, you will do pretty good as a prognosticator.

It is for that reason that preseason projection systems pretty heavily weight recent performance. Most teams just don’t change that much from one year to the next with only a few variations here and there.

While that does work well for a lot of teams, it doesn’t work that great for Florida right now. The 2017 team did not handle Jim McElwain’s exit well. I won’t go so far as to say they quit on the season, but they played well below their potential down the stretch. As a result, the 2017 season’s outcomes are not of much use when projecting the future under a stable coaching regime like what the team has right now. Yet, it still factors heavily in this facet of preseason projections.

And while many of these systems weight recent years more heavily than older years, they do not weight recent games over older games. You can put 2018’s games in any order you want and they will not change their output.

Therefore, they will not account for the fact that Florida’s offense steadily got better through the season. It was not a straight line up, because nothing ever is, but the trend is very clear. The Gator offense at the end of 2018 was much better than the offense at the beginning of 2018.

The 2019 attack could be rough early as the new offensive line figures things out. However it is likely to be on the whole better than the 2018 offense was as long as that line does gel. They return everything at receiver, are still great at running back, have several good options at tight end, and Feleipe Franks should be better than he was last year. His improvements, after all, are a big reason why the offensive trend went up.

The defense shouldn’t be worse than it was last year. Wilson’s return is big like I said in the last section, and Amari Burney brings an ability to cover pass catchers from the linebacker spot that hasn’t existed in years.

There also is the matter of being in the second year of all four systems: offense, defense, special teams, and strength and conditioning. I don’t need to speak much on the first three; most everyone understands the importance of continuity in those systems.

The fourth may or may not be a big deal for a generic new coaching regime, but it is with Florida under Mullen. Strength and conditioning slipped a lot under the previous regime, so much so that AD Scott Stricklin cited it as a factor in his press release announcing McElwain’s firing. Nick Savage’s program showed improvement since last year, but it may be this year when it really comes together.

Take Perine. He has always been an efficient but not an explosive running back. However, he broke off touchdown runs of 74 yards against FSU and 53 yards against Michigan in the season’s final two games. He’d only had two runs of 30+ yards in the entire rest of his career.

Those were not against bad defenses either. That could be a sign that he has added explosiveness thanks to the upgraded strength program, but because it takes time to build real and lasting changes, they didn’t manifest until the end of the season. There are likely to be a number of other such improvements all over the roster.

Overall

A team with a better offense and no worse of a defense should win more games if all else is equal. All else is not equal though. The schedule is likely to be harder, but improvements from being in the second year of the systems should help offset that.

I believe that the numerical projection systems are likely to be underrating Florida in two of their three major categories. The caveats in the third category probably all balance out with it ending up approximately correct, although I could be persuaded that UF is a little overrated in there depending on how the systems work.

Still, it adds up to the Gators probably having lower win count projections than what they should if all of those factors were completely accurate. I have nothing but respect for the people who create these systems, but just like last year, I might add a win to their projections to account for the idiosyncrasies that hurt this year’s Florida team in these kinds of rankings.