Syracuse ended the 2017 season with five straight losses. It allowed 162 points in its last three games. It finished with 4 wins or less for the fourth straight season.
Syracuse ended the 2018 season with a dominant bowl victory, a 10-win season, its first top-25 ranking in nearly two decades and was one snap against Clemson away from clearing the path to an ACC championship.
How in the heck does that happen?
There are a lot of reasons, of course. Dino Babers’ system was finally starting to click. The pass rush blossomed from a unit that had just 16 sacks in 2017 to one with 43 in 2018. Eric Dungey was Superman.
But there was another big reason for the huge turnaround, too, and it might also prove to be a huge red flag for 2019: Turnovers.
Historically, turnovers are an incredibly volatile statistic. Luck plays a huge role, and it’s rare that one team manages to keep a high turnover margin year after year. It’s one of the stats most prone to regression to the mean.
Which leads us to Syracuse.
In 2017, Syracuse’s defense ranked 115th nationally in takeaways with just 12. The Orange were 113th nationally with a -52 margin in points off turnovers. And they were 121st nationally with a turnover margin of -12.
In 2018, Syracuse was 5th in turnover margin (+13), sixth in points off turnovers margin (+65) and third in takeaways (31).
That’s a lot of volatility, even for a statistical metric that’s prone to volatility. In fact, just looking at points off turnovers margin, the standard deviation for change from one year to the next is about 31 points. Syracuse’s change was 117. In fact, Syracuse’s 19 additional takeaways from one year to the next is the most by any Power 5 program in the past decade.
So say it with me: Regression. To. The. Mean.
Historically, the numbers suggest a huge shift back for Syracuse. We took the past 10 seasons of data, calculated the absolute value change in points off turnover margin for every FBS team, then looked at what followed in Year 3 for the biggest outliers.
Put more succinctly: We looked at all teams whose change in points off turnover margin shifted by more than three standard deviations from Year 1 to Year 2, then asked what happened to them in Year 3.
For example:
In 2012, Florida State had a points off turnover margin of +30. That’s right in the expected wheelhouse for any team (i.e. within one standard deviation of the average, which as you’d guess, is about zero.)
In 2013, when FSU won the national championship and dominated the entire season, the Seminoles had a points off turnovers margin of +163. That is insanely good. In fact, it’s the best mark in the past decade — a full 12 points better than the next-best team. Only 6 other FBS teams have even had a mark of +120. In other words, it’s WAY outside the expected range of fluctuation.
So, what happened in 2014? Despite returning the bulk of its talent from that national championship team, having Jameis Winston at QB and a host of future NFL players on defense, and despite going undefeated in the regular season again, FSU’s points off turnover margin dropped to minus-44. That’s a shift from Year 2 to Year 3 of a whopping 207 points. Or, more precisely, over a 14 game season, FSU had to make up 14.8 points per game just based off its shift in turnover scoring. No wonder a team that dominated in 2013 had to endure so many escape acts in 2014.
But here’s the thing: That shift is not uncommon. If a team is way outside the expected norm in points off turnover margin one year, odds are it’ll take a hefty swing back in the other direction the following season.
We looked at 65 teams from 2009-2018 that fit the criteria of falling more than three standard deviations from the mean in points off turnover margin — both good and bad. Of those 65 teams, 15 of them again fell more than three standard deviations from the mean in the opposite direction the following year.
Think about what that means: Teams have about a 95 percent chance (based on our data) of falling inside the meat of our bell curve each year, but 23 percent of these teams beat those odds two years in a row. It’s just that one year they beat the odds on the positive side, and the next on the negative (or vice versa).
That’s not just statistically meaningful. If you’re Syracuse, it should be statistically terrifying.
Looking at the more broad data set, teams that saw an improvement beyond three standard deviations from Year 1 to Year 2 saw a reduction in points off turnover margin of, on average, 65 points. That’s an impact of about 5.2 points per game. Only two of the 26 teams matching our criteria failed to see a regression of less than one standard deviation.
What that essentially means is, if all else is the same in terms of Syracuse’s performance in 2019, it likely is going to be about 5 points per game worse this year, just based on the expected correction in points off turnovers margin.
Now, Syracuse’s wins came by an average of 12.3 points per game last year, and only North Carolina was a particularly close call, so perhaps none of this will matter beyond creating a few more close games that still result in a W.
And Syracuse does return the bulk of that defensive front and talented secondary that accounted for all those takeaways, so maybe the regression falls on the low side of things and the Orange can repeat the success more than most teams have previously.
But history says that the path will be tougher this year. In layman’s terms, we usually say Syracuse won’t be sneaking up on anyone this season, but the data actually says it’s more about what’s being handed to the Orange or, for 2019, what might not be.
*Note: John Casillo of Troy Nunes blog had a follow-up to this looking at some factors working in Syracuse’s favor. Worth a read HERE.
For reference, if you’re the type who might like to wager on team win over/unders, here are the really big outliers in POT margin last year:
Kansas +162
Georgia Southern +124
Syracuse +117
Air Force +105
North Texas +88
Louisville -125
Others worth watching: K-State (-73), South Carolina (-73), Florida (+77), Cincinnati (+75).
by David M. Hale 