It was pretty easy to assume that this college football season would be different. The coronavirus stopped spring football in its tracks, made summer conditioning particularly fraught and altered fall camp practices as well. New coaches of the "head" or "coordinator" varieties had less time than ever to instruct and implement their ideas and systems to their new players.
Plus opt-outs and the potential for positive tests and contact tracing absences meant that coaches had to prepare for a drastic opening up of their respective benches, while writers and fans had to brace for the impact of having less up-to-date information and simply knowing less. So did bookmakers and bettors.
As expected, September was the most volatile first month of a season that we've ever seen. The average absolute-value difference between the spread and the final scoring margin for games featuring two FBS teams was 15.1 points per game -- compared to 13.0 in September 2018, and 12.1 in the Septembers of 2019, 2017 and 2016. Dating to 1978, the highest averages in terms of September errors were 14.1 points per game in 1999 and 13.7 in 1986.
Eighteen of 61 FBS vs. FBS games this September were off by at least 21 points. That's a smaller than normal sample, of course, but that's still nearly one-quarter of games going completely off the rails compared to expectations.
This isn't a case of blowouts getting out of hand, either. The wild swings are coming in games that were expected to be close:
• Games with a spread between 0 and 2.5 points: 19.6 points per game of absolute error (September average since 1978: 12.5)
• Games with a spread between 3 and 6.5 points: 15.1 points per game (11.8 average)
• Games with a spread between 7 and 9.5 points: 19.6 points per game (12.4 average)
• Games with a spread between 10 and 13.5 points: 15.0 points per game (12.6 average)
• Games with a spread between 14 and 20.5 points: 13.1 points per game (12.6 average)
• Games with a spread between 21 and 27.5 points: 15.1 points per game (12.4 average)
• Games with a spread of 28 or more points: 10.8 points per game (12.4 average)
BYU was listed at -1.5 when it throttled Navy 55-3. Tulane and Army were each -3.5 when the Green Wave and Black Knights destroyed Southern Miss and MTSU, respectively, by 42 points apiece. Troy was -2.5 when the Trojans also pummeled poor MTSU 47-14. There have been some other surprises -- Kansas State (+27.5) beating Oklahoma, for instance -- but a lot of the biggest blowouts this year were expected to be tossups.
Meanwhile, life against the spread hasn't been particularly kind to favorites in non-tossup games. (Once again, small sample alert.)
• Favorites between 0 and -2.5: 4-1 against the spread (80% vs. the 51% September average)
• Favorites between -3 and -6.5: 7-8 (47% vs. 47% average)
• Favorites between -7 and -13.5: 5-9 (36% vs. 48% average)
• Favorites between -14 and -20.5: 5-7 (42% vs. 51% average)
• Favorites between -21 and -27.5: 3-6 (33% vs. 53% average)
• Favorites of -28 or higher: 1-5 (17% vs. 49% average)
In all, favorites of three or more points are 21-35 against the spread (38%), and all favorites are 25-36 (41%): 2-3 in Week 1, 5-9 in Week 2, 6-9 in Week 3 and 12-15 in Week 4. There might be a positive trend in there -- 37% in the first two weeks, 40% in the third, 44% in the fourth -- but we'll see. Vegas obviously tends to adapt.
BYU wants to be on this list
Sometimes Vegas can't adapt quickly enough for certain teams, however. We saw it a couple of years ago when, in Bobby Petrino's final year as head coach, Louisville collapsed too rapidly for oddsmakers to catch up and they kept underachieving drastically against the spread all season. And while it's early, we might be seeing this with BYU in the opposite direction.
Kalani Sitake's Cougars began the season with the aforementioned rout of Navy. That was by far the biggest single-game spread outlier of the year, but in their second game they produced the 11th-biggest, blowing out Troy 48-7 as 14.5-point favorites. In two games they've exceeded the spread's expectations by 77 points.
The market is adapting quickly -- Caesars Sportsbook by William Hill lists BYU as a 23.5-point home favorite against Louisiana Tech on Friday -- but even if the Cougars cover by a small amount, they'll be in lofty company when it comes to hot starts.
Using the same set of historic September data, here are the teams that have overachieved the most against the spread out of the gates:
Greatest September spread overachievers since 1978 (3+ games)
• 1983 Nebraska (+28.3 points per game in three games)
• 1987 UTEP (+27.8 in three)
• 1996 Ohio State (+27.0 in three)
• 2017 UCF (+26.8 in three)
• 2016 Ohio State (+25.7 in three)
• 1990 Virginia (+25.5 in four)
• 1982 Boston College (+25.0 in four)
• 2010 Stanford (+24.5 in three)
• 1979 LSU (+24.5 in three)
• 1980 Nebraska (+24.5 in three)
• 2005 USF (+24.0 in three)
• 1994 Florida (+24.0 in three)
• 1978 Texas A&M (+24.0 in three)
• 2016 Troy (+23.7 in three)
• 1999 Cincinnati (+23.5 in three)
• 1980 Florida (+23.3 in three)
• 2012 Texas A&M (+22.3 in three)
• 2012 MTSU (+22.2 in three)
• 1990 Kansas State (+22.2 in three)
• 1979 Alabama (+22.2 in three)
• 1985 Air Force (+22 in four)
This is a list you absolutely want to be a part of. It includes one of Bear Bryant's best Alabama teams (1979), one of the greatest teams ever to not win the national title (1983 Nebraska), and some others that came pretty close to the promised land (1980 Nebraska, 1994 Florida, 1996 Ohio State, 2010 Stanford, 2012 A&M, 2016 Ohio State and, if you're generous with the definition, 1985 Air Force), plus another team that claims a title (2017 UCF). It features some of college football's greatest recent turnaround stories -- the 1987 UTEP team that enjoyed its first winning season in 17 years, the 2012 MTSU and 2016 Troy teams that went from 18 combined losses to 18 combined wins, Bill Snyder's first competitive Kansas State team, etc.
The list also includes almost no one who had anything close to a bad season. The 1999 Cincinnati team that upset top-10 Wisconsin before collapsing to 3-8 is pretty much the exception to the rule. BYU's +38.5 average through two games will obviously come down, but if the Cougars keep the overachievement going for another week or two, they aren't likely to let off the throttle much from there.
This season's top overachievers (min. two games)
• BYU (+38.5 points per game in two games vs. FBS teams)
• Miami (+14.5 in three)
• Army (+14.3 in three)
• Arkansas State (+12.0 in two)
• UAB (+11.5 in two)
• Texas State (+11.4 in four)
• UTEP (+11.3 in two)
• Syracuse (+11.0 in three)
• Notre Dame (+10.8 in two)
• Tulane (+7.3 in three)
A list you don't want to be a part of
On the other side of the ledger, we've had a few teams starting out far more poorly than oddsmakers expected:
This season's biggest underachievers (min. two games)
• Southern Miss (-23.3 points per game in three games)
• Florida State (-23.3 in two)
• MTSU (-21.7 in three)
• Navy (-21.0 in two)
• ULM (-19.5 in three)
• Kansas (-18.5 in two)
• Iowa State (-15.5 in two)
• Western Kentucky (-11.0 in two)
• Duke (-10.2 in three)
• Georgia Tech (-9.0 in three)
You can see some small-sample effects in there with teams like Navy, but the top, er, bottom three teams have looked pretty bad multiple times. MTSU might have begun a rebound in Week 4 -- the Blue Raiders still lost to what likely is a pretty bad UTSA team by two points, but they were six-point underdogs and were unlucky to lose -- but they have clearly not been very good. Neither has Southern Miss (which sandwiched an unlucky loss to Louisiana Tech between a bad performance against South Alabama and a horrid one against Tulane) or Florida State.
The Seminoles are off the boards this week, hosting FCS opponent Jacksonville State, but Southern Miss is somehow only a 1.5-point road underdog against North Texas. If there's any consolation for these teams, it's that there are a few more examples of teams turning things around after poor ATS starts -- at least as long as you haven't started too poorly:
Biggest September spread underachievers since 1978 (3+ games)
• 2016 Bowling Green (-37.0 points per game in three games)
• 2008 Washington State (-30.5 in four)
• 2012 Arkansas (-29.4 in four)
• 1997 Texas (-27.7 in three)
• 1991 Michigan State (-27.3 in three)
• 2017 Missouri (-25.8 in three)
• 1999 SMU (-25.5 in three)
• 2019 Bowling Green (-24.3 in three)
• 2008 Idaho (-23.8 in four)
• 2008 Virginia (-23.5 in three)
• 2010 New Mexico (-23.3 in four)
• 1999 Arizona (-23.1 in five)
• 1997 Louisiana (-22.6 in four)
• 2004 North Texas (-22.6 in four)
• 1996 Clemson (-22.3 in three)
• 2014 SMU (-22.3 in four)
• 1980 Cal (-22.2 in three)
• 2000 Louisiana Tech (-22.2 in three)
• 2008 Rutgers (-22.2 in three)
The bad news? Of these 19 teams, only four rebounded to finish with a winning record, and only one of the 12 teams at minus-23 PPG or worse did. A lot of the teams here went from mediocre to horrid (1980 Cal, 1997 Louisiana, 2000 Louisiana Tech, 2008 Washington State, 2014 SMU) or from good to mediocre (1991 Michigan State, 1996 Clemson, 2008 Virginia), and a few utterly collapsed under the weight of high expectations (1997 Texas, 1999 Arizona, 2012 Arkansas).
Florida State can perhaps draw inspiration from the 2017 Missouri team that looked completely unprepared at the start of the season but rebounded from 1-5 to win six in a row, and Southern Miss could look to the 2004 North Texas team that started 0-4 and then won seven straight. But Week 5 offers them opportunities to steer out of their current skids, as MTSU may or may not have begun to do in Week 4.