clock menu more-arrow no yes

Filed under:

Predicting Florida’s 2017 season by win shares: Fall practice

New, 88 comments

We’re going to do this weekly and diligently this year.

NCAA Football: Vanderbilt at Florida Kim Klement-USA TODAY Sports

One of the things we’ve grown fondest of doing while making our predictions about the Florida Gators around these parts is using win shares to do it.

As longtime readers of Alligator Army know, the use of win shares has been part of our annual predictions of every season dating back to 2012; in 2016, I updated the format by trying to do it weekly, and ... well, I tried, y’all. As ever, I think win shares are an interesting, somewhat rigorous means of determining just how much success we really think the Gators are likely to have in a given year based on the confidence we have in each given Saturday’s game.

The way it works is simple: Come up with your percentage change for a team to win a given game, however you will, and convert it to a decimal value. (100 percent is 1.00, for example.) Then do that for every game, and tally the numbers to come up with a predicted win total for the season.

Thus, here is my first go at 2017 win shares for Florida, on the day of the Gators’ season opener.

Florida Gators Win Shares Predictions (August 9, 2017)

Opponent Win Share
Opponent Win Share
vs. Michigan 0.50
Northern Colorado 1.00
Tennessee 0.70
at Kentucky 0.75
Vanderbilt 0.85
LSU 0.55
Texas A&M 0.65
vs. Georgia 0.65
at Missouri 0.75
at South Carolina 0.75
UAB 0.95
Florida State 0.40
Total 8.50

Some notes:

  • Michigan is favored to beat Florida in Dallas by a field goal, so, immediately, my win shares diverge from conventional wisdom by considering that game a true toss-up. But Michigan lost even more on defense than Florida, and Michigan didn’t have the months-long rash of season-ending injuries that Florida did, so Michigan’s defenders haven’t spent as much time on the field for consequential snaps as Gators like Kylan Johnson, Vosean Joseph, and Chauncey Gardner have. Of course, Michigan also lost to the same Iowa team that Florida waxed in the Outback Bowl, so asking “Which team drops off more from 2016, Florida or Michigan?” isn’t necessarily asking the right question. Regardless — and, obviously, we’ll talk more about this in next few weeks — I am fairly confident Florida will be competitive in that game at worst, and the Gators doing that will go a long way toward making this season successful or a struggle.
  • I have vacillated over time on whether I should really give 1.00 values for win shares, even in games against clearly overmatched teams. But I did it last year for both Massachusetts and North Texas, both thought of as truly awful FBS teams prior to the season, and Northern Colorado’s a middling FCS team. And that game happens early enough in the year that Florida shouldn’t be sideswiped by catastrophic injuries like it was prior to that infamous Georgia Southern loss — while, as you see, the UAB game happens late enough in the year that I’m baking in some room for “Hey, anything could happen” in that value.
  • Florida’s favored over Tennessee in that game at home, and Florida is almost certain to come out for that game fired up to re-establish dominance over the Vols after collapsing in Knoxville in 2016. Given that Joshua Dobbs isn’t around to give Florida’s defense fits, I like the Gators’ chances of doing that.
  • Home-field advantage is probably worth about 10 to 20 percent of confidence for me, depending largely on the home field. Traveling to Kentucky in September, for example, is not the same as traveling to Missouri in November for what could conceivably be a snow game. But that doesn’t make sense for the Kentucky and Vanderbilt values being as they are — either Kentucky should be lower or Vandy should be higher, right? — until one realizes that Florida has one loss in Lexington since 1976. Florida has lost more home games to Kentucky than it has road games in the last 40 seasons. This is another year in which Kentucky fans who saw their team knock off Louisville and stand and trade with Georgia and Tennessee late in 2016 can dream of finally upending Florida, but Florida also drummed that same Kentucky team last year, and the Gators did that with Luke Del Rio at quarterback.
  • That three-week span of October with LSU and Texas A&M coming to The Swamp prior to the trip to Jacksonville is obviously the key stretch of this season for Florida, and one that could spell a 10-win season or a five-loss season if either extreme outcome happens. My win shares impute that going 2-1 over those games would be doing slightly better than expected (0.55 + 0.65 + 0.65 = 1.85), but also that Florida may well be favored in all three games. Certainly, an October spent entirely in the Sunshine State seems far less daunting than one cleaved into pre- and post-trip to LSU.
  • Traveling to Missouri and South Carolina in consecutive weeks might be new not just for Florida but for any SEC team. Doing so in November, when the weather could be cool or cold in both Columbias, makes that an even trickier back-to-back. But Missouri was really bad in 2016, especially defensively, and South Carolina was actually worse per S&P+ — thanks to, and you’re not going to believe this, an anemic offense — despite the spark Jake Bentley provided as the season waned. But while Florida nose-dived on offense in November 2016, the Gators’ best quarterback from last year — the one who wasn’t even healthy enough to play in November — is now likely to be the third-stringer, and the offense (and defense) should really be at whatever peak level they will reach even with normal injury luck. If we aren’t pretty confident about those road trips when they actually roll around, I think it’s likely something will have gone very wrong.
  • UAB being a football program again after a two-year hiatus will be fascinating this fall. I don’t buy that UAB will be any good, though: Bill Connelly’s S&P+ projects the Blazers will be not just No. 130 of 130 teams, but about a touchdown worse than No. 129 Texas State, and projects a margin of 49.5 points between the Gators and Blazers, which might be the most lopsided preseason prediction it makes. The 0.95 instead of 1.00 is about injuries and so forth, not any currently extant worry that UAB will be a threat to Florida.
  • Florida State is favored over Florida right now, and I’m skeptical about predicting any Florida team will beat Florida State until it actually happens. Florida will likely know whether it is ticketed for the SEC Championship Game prior to this game, though, and could have the confidence derived from reaching the 10-win plateau, too — and FSU will not have Dalvin Cook, as effective a Gator-slayer as any ‘Nole has ever been. There’s copious young talent on both rosters, but any reasonable analysis would give Deondre Francois the edge over Florida’s QB — whether Feleipe Franks, Malik Zaire, or Del Rio — at the moment, and that advantage is worth something in my estimation. Expect this value to shade up or down more than any other this season, but also expect that it might take a lot to get it above 0.50.

All that said: These are my win shares for Florida’s 2017 season from where we sit at the end of the first week of fall practice, a sort of pre-preseason representation of my thinking. I’ll review and revise — if necessary — in the last week of August, posting a final preseason win shares prediction on Wednesday, August 30. From then on, I’ll also post updated win shares predictions each Wednesday — Win Shares Wednesdays! — for the rest of the year.

And, as always, I encourage you to post your own win shares and quibble with mine in the comments. Are you more bullish than I am? More bearish? Do you see a game I’m completely misreading with clearer eyes? Let me know.