A nerd's look at predicting the SEC, Part I: Record in close games
Advanced statistical metrics, or analytics, entered the mainstream culture years ago.
The use of carefully-culled and interpreted data is generally accepted as a more accurate predictor than the “eye test” or rudimentary stats like Earned Run Average in baseball.
From the Hollywood movie “Moneyball” about Billy Bean, MLB and NBA executives hired solely due to their statistical understanding to big media attention on sites like ESPN, the analytics movement is an “in your face” part of our sports culture.
Its application within the game of college football is important. It admittedly lags behind sports like MLB (162-game regular season) and the NBA (82-game regular season).
Even the NFL has an advantage in terms of the use of analytics, because a) forward-thinking franchises devote more monetary resources toward it than most college programs, b) public websites like Pro Football Focus and Football Outsiders collectively produce more analytics-related content about the NFL than what’s out there for college football, c) the level of competition is more even in the NFL and there are no Auburn-Samford matchups to throw off the data, and d) the NFL features a 16-game regular season as opposed to the 12 games in college.
More plot points and more data entries equate to stronger conclusions. But there still are a number of statistical measures within the college game that give us a greater understanding of which teams are lucky, which teams are great and which teams are both. Performance measures that account for the outcome of every play rather than every game are much more accurate, as the number of plot points can go from 13 to 1,000 quickly. There are a few of those in college football, such as the F+/- metric produced by Football Outsiders.
We’ll spend this week looking at a few of the game-by-game, outcome-related numbers that remain useful, even in college football.
The first of those, which we’ll look at today, pertains to a team’s win-loss record in close games. We’ll define a close game as a margin of eight or fewer points, or a one-possession margin.
There’s one surefire way to finish a season with a good record in college football, and that’s to get to the fourth quarter with all of your starters on the bench as your reserves finish out a blowout victory.
History has proven that even for teams with unique characteristics, outside of quarterbacks who are elite at running the no-huddle, two-minute offense, there are very few repeatable advantages in close fourth-quarter games. The outcome is more akin to a coin flip.
Here’s where the term “regression to the mean” becomes meaningful.
As Grantland’s Bill Barnwell wrote, regression to the mean “suggests that a prior event has no useful insight into predicting whether a random event will happen again in future tries, and that the most likely outcome over the next number of trials will be an average expectation.”
In other words, imagine you flip a coin and it lands on heads 10 consecutive times. Regression toward the mean would see the next 10 coin flips land on heads five times and tails five times. The coin may land on heads every time, or it may land on tails every time, but the outcome is random.
Take that coin and flip it 100,000 times, though, as opposed to 10, and it becomes more and more likely that you’ll see something close to 50,000 outcomes of heads and 50,000 of tails.
To examine how this applies to SEC football in close games, here’s a look at the record of all 14 teams in such games the last three seasons.
The SEC as a whole is seven games above .500 in close games in the last three years, including 33-33 in 2014.
But let’s parse it a little further to illustrate regression to the mean.
In ’12 and ’13, there were eight different cases where an SEC team finished at least two games below .500 in close games. The combined record of those teams was 4-23, representing a .148 winning percentage. If there was some tangible reason or rule dictating that those teams weren’t very good in one-possession games, one would expect similar results in the following season. Instead, those same teams finished a combined 15-21 in the year after struggling in close games, resulting in a .417 winning percentage.
Similarly, specific SEC teams finished at least two games above .500 in close games in nine cases during the ’12 and ’13 seasons. Those teams combined for a 35-11 record (.761) in close games. The following season, those same teams finished a combined 23-20 (.535) in close games.
When games reach the last half of the fourth quarter, and all the previous possessions haven’t led to any separation, and a final one or two possessions are going to make the difference in the outcome, the result over time is more or less random.
HOW IT APPLIES TO THE SEC IN 2015
Examining each team’s record in close games successfully projected a decline for Auburn in ’14.
The Tigers, a so-called “team of destiny” in coach Gus Malzahn’s first season as the head man, went an astonishing 6-0 in one-possession games during the regular season, seizing a surprise SEC championship. Last year Auburn finished 8-5, losing to Texas A&M, 41-38, dropping the Outback Bowl to Wisconsin, 34-31 in overtime, and very, very nearly losing to Ole Miss on the road before pulling out a 35-31 victory on the Laquon Treadwell injury play.
After finishing a combined 7-2 in close games during the ’12 and ’13 seasons, both 11-win campaigns for the South Carolina Gamecocks, coach Steve Spurrier’s program dropped four close games last year. With a frenetic offense and leaky defense, the Gamecocks got involved in a lot of high-scoring, race-you-to-42-points contests. South Carolina finished 3-4 in those games, portending the team’s four-win decline and 7-6 overall record.
The Gamecocks were better in ’12 and ’13, but in those seasons they got a little lucky. In ’14, the team got worse, and its luck wasn’t as good.
So, who can we expect to benefit and suffer during the 2015 season based on this rule?
“Lucky” SEC Teams
Auburn (9-3 the last two years)
Mississippi State (6-2 the last two years)
Missouri (4-1 last year)
Vanderbilt (8-3 the last three years)
“Unlucky” SEC Teams
Arkansas (1-11 the last three years)
Florida (3-7 the last two years)
Kentucky (1-6 the last three years)
Missouri fans won’t like to hear this, but the team utilized some luck in pulling out four of its five close games last season, which helped the Tigers win an SEC East title. That’s in stark contrast to the ’13 Missouri Tigers, which was such a dominant team that it only faced two close games all season, going 1-1 in those contests.
Mississippi State also has outperformed the expected outcome in close games in the last two years, winning 75 percent of the time. That’s yet another reason why it will be tough for the Bulldogs to repeat their temporary, impressive run to No. 1 in the country last season.
A pair of SEC East teams that have struggled the last few years, the Gators and Wildcats, may inch up the standings by winning an additional game or two in ’15, as both programs are due for some better luck. (Ironically, one of Florida’s few wins in close games was a 36-30 win against Kentucky in triple overtime.)
That brings us to Arkansas, which deserves its own section.
THE BRET BIELEMA FACTOR
The Razorbacks either are way overdue for some good luck in close games, or represent the rare team whose style makes it more likely that the team drops a larger percentage of those contests.
Arkansas coach Bret Bielema has instituted a somewhat unique, against-the-grain style in Fayetteville, as the team relies on large offensive linemen, a physical, punishing ground game and safe throws to a cadre of tight ends.
Bielema’s recent record in close games, including some of his time at Wisconsin, is shocking: 3-16 since 2011.
The Hogs haven’t won a single close game since Bielema became coach, including an 0-4 record last year. The 2014 Arkansas Razorbacks (7-6) won by an average of 31.7 points per game and lost by an average of 9.8 points per game, including a 24-point defeat at Auburn to open the season.
But Bielema has been very successful in close games running the same style of offense, albeit in the Big Ten. His Wisconsin teams went 18-7 in close games from 2006-10, so maybe Bielema was just due for a correction with his good luck.
Bielema’s teams in close games
2014: 0-4 (Arkansas)
2013: 0-3 (Arkansas)
2012: 2-6 (Wisconsin)
2011: 1-3 (Wisconsin)
2010: 2-1 (Wisconsin)
2009: 6-1 (Wisconsin)
2008: 3-3 (Wisconsin)
2007: 4-2 (Wisconsin)
2006: 3-0 (Wisconsin)
I examined every Arkansas offensive snap in the fourth quarter or overtime of close games during the last two seasons, and the analysis is shocking.
There are three main reasons that the Razorbacks have failed to win those games.
REASON 1: BALL SECURITY
The Razorbacks’ turnovers in close games amounts to sadistic comedy, or perhaps masochism for Hogs fans. In seven such games (we’re talking seven quarters and two overtimes), Arkansas has thrown three interceptions and lost three fumbles.
Take a look at the comedy of errors the Arkansas offense committed in six of the seven losses. (The Razorbacks blew a 17-point late-third quarter lead to Rutgers in the other game.)
24-17 OT loss to Mississippi State: Alex Collins loses a fumble in the red zone in regulation, Brandon Allen throws fourth-down INT in OT
31-27 loss to LSU: Strip sack of Brandon Allen, Tigers run out the clock
35-28 OT loss to Texas A&M: Missed 44-yard FG in regulation, stuffed on fourth-and-1 in OT
14-13 loss to Alabama: Brandon Allen throws final-possession INT
17-10 loss to Mississippi State: Missed 42-yard FG in regulation, Brandon Allen red zone INT on final possession
21-14 loss to Missouri: Alex Collins lost fumble on final possession
REASON 2: INEFFECTIVE PASSING GAME
The Razorbacks thrive on a powerful running game, producing a pair of 1,000-yard rushers in ’14. The team pairs that focal point with a safe play-action passing game that prevents turnovers and takes advantage of the team’s plus talent at tight end.
But late in close games, especially when trailing, the team isn’t well equipped to throw down the field or move the ball quickly.
Here are the QB numbers Arkansas has compiled in its seven close games (all losses) under Bielema: 28-of-66 (42.4 percent completion rate), 293 yards (4.4 yards per attempt), 0 TDs, 3 INTs, 3 sacks.
Tight end Hunter Henry is responsible for 10 of the team’s 14 passing first downs during that period. Another tight end, Jeremy Sprinkle, claims one of the others: a 44-yard reception that represents the team’s biggest play in the fourth quarter or overtime of those losses.
Defenses know exactly what to do and whom to blanket when Arkansas needs points in a bad way late in close games. As a result, the Razorbacks have not scored a single point in the fourth quarter or overtime of close games during the Bielema era.
REASON 3: POOR EXECUTION
Arkansas just couldn’t do anything right late in close games the last two years.
The team attempted two field goals in the fourth quarter of those games — of 44 and 42 yards — and missed both.
Against Texas A&M, Arkansas failed to convert a fourth-and-1 on a running play, which is inexcusable for a team as powerful on the ground as the Razorbacks against one of the worst run defenses in recent SEC history.
Arkansas stands to improve its ball security, which may be somewhat of a fluke, especially the fumbles. But it does make sense that the team’s offense would turn the ball over at a higher rate when facing poor down-and-distance situations and attempting to morph into something that isn’t comfortable for the personnel.
The poor execution should revert to the mean as well. The team isn’t going to miss every medium-length field goal it attempts late in games and won’t fail to convert on fourth-and-1 very often.
The passing game is the biggest concern. At the highest levels of football, good teams will make you beat them left-handed. That’s what’s happened to the Razorbacks against SEC opponents the last two years. If teams get them down late in games, Arkansas hasn’t been able to win with its weakness, the downfield passing game.
Fifth-year senior Brandon Allen has gotten hype as one of the SEC’s most experienced quarterbacks and for a sharp-looking spring game. Arkansas also has introduced a new coordinator in Dan Enos. But the personnel and playbook are essentially the same. And unless teams let Henry, an All-SEC type tight end, beat them in the fourth quarter, the Razorbacks don’t have much ammunition in the passing game.
History dictates that Arkansas is due for some luck in close games, and in a big way. But I’m not so sure that the team isn’t the rare true sub-.500 program in close games due to its outlier style of play.