While this is a week to see the talent that Butch Jones and his staff is bringing to Tennessee and look ahead at what the future might bring (and seriously, check out our NSD prep articles, they are great), SBNation's #1 stat-man Bill Connelly has been using the offseason to look back on 2014, taking a deeper look into some numbers and seeing what we can learn from them. One of those numbers ended up being a win expectation based purely on performance. If a team replayed each game with a performance of of similar quality, what are the chances they would win (or lose) again? Hunter put up a fantastic post about this on Monday, taking a look at the numbers and asking how Tennessee's win total compared to their overall level of play. If you haven't read that, do it now.
But how many you won based on your level of play is not the only interesting thing to examine when looking back on 2014. We can also ask how many games you win compared to how many you're expected to win. And where is the best place to look for expectations? Vegas, naturally. Based on the 2014 point spreads and charts that take past data and correlate lines to win probabilities, we can look at how Tennessee played compared to both preseason and in-season expectations.
How did Tennessee do compared to preseason expectations?
|Arkansas State||?||.85 (est.)|
There were three games that didn't have preseason lines, but we can make educated guesses in those cases. We do have information for the other nine games, so making guesses for Arkansas State, Chattanooga, and Kentucky allows us to come up with a preseason expected value of wins: 5.17. And using the little auto-combinatorics page that KidBourbon found for us back in the summer, we see that if we take those expected win percentages as fact, Tennessee's chances of reaching at least their actual six-win mark were just 40%.
It's safe to say that the Vols looked quite good compared to preseason expectations in 2014--they finished nearly a full game ahead of what was expected by the biggest group of objective observers in football. But preseason expectations are notoriously unreliable. So we can also look at the closing line of all 13 games from the 2014 season and see how Tennessee played each week compared to the expectation for that week. This number will be different than the preseason number.
Now you may object that if in-season expectations are higher than preseason expectations, that just means that Tennessee has played well enough to raise expectations. They shouldn't be punished for that. And similarly, if in-season expectations are lower, that just means that Tennessee played poorly enough to lower expectations, and that shouldn't excuse a poor performance relative to preseason expectations.
To an extent, you'd be right to raise such a concern. But the Vols don't play in a vacuum, and their play isn't the only thing that affects the in-season expectations. When you see Tennessee go from a preseason underdog to an in-season favorite against Missouri, that's because the Vols are playing well. But when the South Carolina and Vanderbilt lines dropped ten points, that's not all on Tennessee. South Carolina just didn't turn out to be the top ten team they were projected to be, and a Vanderbilt team that expected to be in bowl contention put those thoughts to bed on opening night when they were walloped by Temple.
So how did the Vols compare to in-season expectations? Let's find out:
You'll notice that the numbers stay relatively stable for the first month of the season. That's to be expected. We then have sizable increases in odds to beat Florida, South Carolina, Missouri, and Vanderbilt and a slight but noticeable decrease in odds to beat Ole Miss. This should dovetail nicely with your memory of this season. Tennessee played an Ole Miss team that was riding high and ranked in the top five and played Florida, South Carolina, and Vanderbilt teams that had vastly underperformed. Missouri might look a bit odd there, as they didn't underperform expectations this season, and I think it's safe to say that the dramatic change in odds to beat the Tigers was based more on people being impressed with Tennessee's defense and newfound quarterback than it was Missouri looking significantly worse than anticipated. This is the same reason the Kentucky line looked about the same as people might've expected in the preseason despite the fact that Kentucky had a much better season than anticipated. The Cats were clearly improved, but so were the Vols.
Adding together the regular season win probabilities gets us an in-season expected value of wins: 5.97. Adding the bowl game brings that number to 6.54. Using our combinatorics page with the in-season numbers gets us a 64% chance that the Vols would win at least six games in the regular season and a 51% chance that the Vols would win seven total, assuming the actual bowl matchup against Iowa. Basically, the Vols were dead on their in-season expectation in the regular season and came out flying in the bowl.
It's easy to get wrapped up in what people in Knoxville expect and tell our own story about the season. An optimist might tell the story of how Tennessee wasn't given much of a chance to make a bowl and how the Vols overcame adversity to reach the postseason for the first time since 2010, riding a defense that was much improved and discovering an offense late in the season that seemed to score at will in three of the last five games. They might continue by looking at narrow losses to Florida and Georgia and talk about how Tennessee was just a few points from reaching eight regular season wins and blowing preseason expectations out of the water.
A pessimist might tell the story of how the SEC East hit its lowest point in years and how the Vols weren't able to take advantage. Of how Tennessee blew a winnable game in Athens and choked away their best chance to beat Florida in ten years. They might continue by looking at an overtime win in Columbia and a far too close game in Nashville and talk about how Tennessee was a Cam Sutton punt return and a pair of overtimes--one real overtime, one hypothetical--away from its first eight-loss season in school history.
But when we look at what outsiders think, we can get a less biased picture of how the season went. And how it went is that Tennessee looked good--but not unbelievably so--compared to preseason projections. Nobody expected the defense to be what it was, and making a bowl was a pleasant overachievement. But it wasn't the kind of massive overperformance of expectations that more stable programs have seen in year two.
And what about given the state of the SEC East? Tennessee looked a little better than expected. The difference between 5.97 and 6 wins means approximately nothing, but we have to keep in mind that some of those in-season expectations were Tennessee-driven, not the-SEC-East-sky-is-falling-driven. Given the rough seasons that Florida, South Carolina, and Vanderbilt had, Tennessee should've expected to win six games--but not easily. Given that and the change in expectations that came from a few unexpectedly good performances, Tennessee was able to keep pace with the expectations it generated in the regular season and then leave people with a positive aftertaste in a bowl shellacking of Iowa.
Despite what the optimist may remember as this season gets farther away, this wasn't a team that improved so quickly that expectations couldn't keep pace with them. They improved, but as improvements go, it wasn't on the overly dramatic side. But despite what the pessimists may remember, this also wasn't a team that couldn't meet the increased expectations as the season wore on. They struggled in some winnable games, but they played well in some losable games too.
If the Vols break through next season and win nine or ten games, this season will be seen as the springboard to bigger and better things, the season that showed that Tennessee was on its way back. If the Vols falter to another 7-6 year, this season will be remembered less for the improvement and more for the failures to break through. But ultimately, 2014 was a season of steady, generally predictable improvement. It was a moderate success, no more, no less.
Now let's get some recruits and make next season one for the ages. Go Vols.