Photo Credit: John S. Peterson

Fumbles Are Dumb

July 12, 2019

Fumbles are dumb.

John Heisman—namesake of the famous trophy, football pioneer and coach—had the best quote about fumbles: “Better to have died as a small boy than to fumble this football.” This was reportedly part of Heisman’s annual preseason speech to his teams (he coached at eight different schools including Auburn, Clemson and Georgia Tech).

It’s almost impossible at this point to know if he actually said that and doesn’t matter anyway. It’s part of football folklore now, such a part of the philosophical firmament of the game that you can buy that quote on a poster featuring a photo of the sun setting on a tropical island for less than $14 from And I have. Because, real or not, that’s a good quote and, with that particular photo, it makes little sense.

Much like fumbles make little sense. The best thing ever said about them is also a little dumb by definition. Has to be. It’s about fumbles.

Proving that fumbles are dumb isn’t particularly hard. They occur with almost equal frequency anywhere on the field. They happen to teams that are ahead and teams that are behind. Fumbles befall teams with coaching staffs that really stress ball security and drill it every day, a group that includes all coaching staffs. While there’s variance from team to team on a year-to-year basis, every year the recovery rate for all fumbles in college football is just a fraction of a percentage point or two away from 50%; half stay with the team that fumbled and half go to the other team. I don’t have all the data I need to prove this, but I’d be willing to bet my most prized possession, the Heisman poster about fumbles on a tropical island, that the overall recovery rate for every fumble in the sport’s history is 50%. Practically speaking, witnessing a fumble is like watching the game stop for a coin flip to determine which team will have the ball on the next play.

Long before overstating the importance of football had replaced baseball as America’s pastime, Heisman was ahead of the game by making fumbles a matter of life and death. It was a nice try, but I assume, over the course of his career, Heisman’s teams recovered about 50% of their fumbles, too. Just no way around it, really, with a large enough sample size.

And this, at least in part, is why opening drives are important.

Scott Frost is merely average at winning coin flips as a head coach. Coach enough years and it’s impossible for it to be any other way, but over 38 games and three seasons Frost’s team has won the toss 19 times. There was a somewhat unlikely run at the start of the 2017 season when Central Florida lost six consecutive coin flips to start the season, but the gods of probability will have their way eventually.

Things get interesting, however, after the coin toss, that mundane thing we do just so we can get the spectacle started on fall Saturdays.

When Frost has had the choice to receive the kickoff or defer the choice to the second half, he has chosen to put his offense on the field first nearly 74% of the time. It’s the opposite of what could be considered basic coin-toss strategy for football coaches.

Believe it or not, there’s no great data source for the results of coin flips in college football games over many years. At least not one I have access to. But in a 2008 story USA Today noted that “college coaches defer more than 90% of the time.” USA Today was running a story about coin flips in 2008 because the NFL had just changed its rules to allow the coin-toss winner to defer to the second half. Just 32% of teams did over the first three seasons after the change but the percentage climbed to 49% in 2012 and 83% headed into the 2018 season, according to ESPN. I think you can safely say the average coach chooses to defer at least as often as Frost chooses to receive.

Football coaches copying one another is a time-honored tradition. I assume most prefer to defer simply because most others prefer to defer and thus it seems less risky. For those that do, the reasoning most often cited in the few articles written on this topic is the chance to “double up,” i.e. score right before halftime and then score again on the first drive of the second half. New England Patriots head coach Bill Belichick doesn’t really express a preference for any strategy but the numbers show this is his preferred choice and it’s probably not coincidental that this has become the dominant strategy in the NFL.

To me, the non-football coach, deferring––unless there are undeniable reasons to do so, like a stiff wind––seems like little more than procrastination. The goal of the game is to score more points than the other guy and to defer is to say, “We’ll get to that a little later.” That’s great news for those that are happy to go first.

Like Frost.

Not only has he taken the ball 14-of-19 times when given the choice, opposing coaches have given his team the ball at the start of the game 16 times. In 30-of-38 games as a head coach, Frost’s offense has had the ball to start the game. By preferring the less common choice, Frost gets what he wants most of the time.

If you have an offense that’s pretty good at scoring points, that can be an advantage. Over the last three years of college football, the team that has scored first has won 67% of the time. This is one of those stats that’s easy to dismiss as happenstance. Can one score, presumably early in a 60-minute game, really make that much of a difference? All of the potential reasons for deferring—the chance for back-to-back possessions, wanting to get a feel for the game, weather, tradition, indifference—seem to make more intuitive sense than citing that, on the whole, the team that scores first has won two-thirds of the games over the past three seasons.

That’s all games. Of course, all games and teams are not created equal. Getting on the board first as a heavy underdog may simply be leveling the playing field some. For the heavy favorite the first score might be a nearly fatal blow. Alabama is already almost impossible to beat but it becomes nearly immortal when scoring first. The Tide are 29-1 over the last three seasons when scoring first. They’re 12-2 when they don’t over that same span, still not great odds but better than you’ll get when Alabama beats you to the punch.

Scoring first has more power than it’s given credit for. Mike Riley was a .500 head coach overall at Nebraska, but he was 14-5 (.737) when the Huskers scored first. Bo Pelini had a .713 winning percentage at Nebraska, but was 49-9 (.845) when scoring first. Frost’s numbers aren’t quite as stark. Yet.

He is 23-15 (.605) overall and 14-8 (.636) when scoring first, a little short of the national average. But his three-year career has been one of extremes. Frost faced two significant rebuilds in his two Year 1s (2016, 2018) and his only Year 2 team (2017) beat everyone regardless of the first score. UCF did score first in eight of 13 games that year, however.

The great thing about football is that one thing almost never wins a football game on its own, but coaches still have to decide what they’re going to emphasize, how they’re going to plan to win. Despite what football folklore may lead you to believe, trying for the first score is no less valid a path to victory than a more time-worn truism—field position. A 1- to 6-yard edge in average starting field position is only producing the winner 55-to-60% of the time.

A 7-0 lead 90 seconds into the game is worth more than that. Why?

There are probably a lot of psychological reasons an early lead could have an impact on the game. To go down the field and score the first chance you get is something of a mission statement, a punch in the mouth. The team that trails is immediately on the defensive when its offense gets the ball, perhaps consciously but definitely subconsciously.

“Everything that happens in a football game takes its meaning from the interplay of time, score and field position,” wrote the authors—Bob Carroll, Pete Palmer and John Thorn—of The Hidden Game of Football, a book published in 1988. A football coach only sort of has control over two of those dimensions.

A team can control how it chews clock or tries to conserve time when it’s on offense, but at the end of the day 60 minutes is 60 minutes. Score, of course, is the one thing that matters most and is largely determined by actual skill, decision making and luck. Field position, a medley of all three phases of the game, can influence score.

In this three-dimensional view of football, scoring first alters two of those dimensions—time and score. Perhaps the easiest way to view this is if you were a football coach would you rather play a 60-minute game that starts 0-0 or a 58-minute game where you’re in a 7-0 hole? Nobody is choosing the latter but this is how a first score is influencing the game and, unless you take the ball first, it’s the risk a deferring team is taking.

The win probability calculators that have become popular of late are based on those factors mentioned above—how much time is left, what’s the score, which team has possession and where—and modeled off of actual results. In an evenly matched game Team A may have a 51.8% chance of winning as soon as it elects to take the opening kickoff. That’s the power of possession. On the first drive, Team A drives 75 yards in three minutes and scores a touchdown to boost its win probability to 67%. Team B does the same on its first offensive drive—a three-minute, 75-yard touchdown drive—and gets its win probability back to 48%. It’s not 50-50 because Team A now has the ball and time has elapsed. If the teams continue trading the exact same touchdown drives in the first half, Team A’s win probability will never fall below 51.8% and Team B’s will never rise above 48.2%. Team B’s only chance to flip the odds in its favor is to get a lead and its only chance to do that is by scoring on the last possession of the first half and the first possession of the second. It has to “double up.”

If Team B does take the first lead of the second half, it gets a big bump in win probability because half the game is already gone. A late lead is, of course, better than an early lead.

Of course, this sort of all-touchdown game doesn’t really exist (outside of maybe the Big 12), but it’s useful to understanding why a lead has more value than just what shows on the scoreboard. To win, a team has to score. To get the lead is to unlock the potential for a win and open up the possibility of never trailing. Every lead earned exerts a pressure that is easy to miss when your team is “only down a field goal” with “plenty of time left,” but every lead has value and its value increases as time decreases.

Simple enough, but is that alone enough to make pursuing the first score a priority when given the choice at the start of a game?

If it’s not, what happens with turnovers might contribute to the argument.

Turnovers, as a whole, are not dumb. They can decide football games.

Over the past three years the team that has won turnover margin has won about 73% of the time. Since 2007, Nebraska is 43-3 when winning the turnover battle and 25-50 when losing it. A truly bizarre thing happened last November—the Huskers lost a game, for the first time since 2003, while being plus-two in turnover margin. Nebraska had three takeaways against Ohio State to just one giveaway but it wasn’t quite enough. (Nebraska scored first, on the first overall drive of the game, too.)

Numbers like that make you wonder if maybe John Heisman didn’t go far enough in his efforts to eliminate fumbles. “Better to have died a small boy and been condemned to hell than to fumble this football.”

Then you remember that fumbles are dumb (luck) and there really isn’t anything anyone can do. Some offenses, particularly option offenses, are a little more fumble-prone, but the alternative is to not run the offense you want to run.

Interceptions, though? Interceptions are interesting. There’s a really easy way to avoid those—don’t pass. Isn’t that the same non-alternative alternative as giving up the option to avoid fumbles? Not quite. Picture almost any offense you can think of, from air raid to flexbone and everything in between, and then ask yourself when does this offense pass the least?

When it’s ahead.

This is a key distinction. To play football is to fumble, but to throw interceptions is only to pass.

While 57.6% of all turnovers since 2016 have been committed by the offense that trails, it’s because of interceptions, not fumbles. Over that span, offenses have run 45.4% of their plays while trailing and 47% of fumbles have happened while behind. Based on that, you might expect about 45% of interceptions to be thrown when trailing. But they aren’t. Of all the interceptions between 2016 and 2018, 62.8% of them have been thrown by the team that trails. And, no, this isn’t simply a matter of a team throwing more when its behind. Not entirely. Trailing teams threw 51.1% of the passes those three seasons, still less than the interception rate.

To really show the impact score is having on turnovers, let’s break this into smaller groups based on the size of the lead—leading by 15-plus points, 8-to-14 points, 1-to-7 points, tied and then trailing by the same increments. That’s essentially a tied game a one-score game, a two-score game and anything more than two scores. The fumble rates (fumbles lost divided by total plays) at each of these game states remain nearly the same; they ranged from 0.55% to 0.66% but there wasn’t a gradual increase as you move from having a big lead to a big deficit. But that’s what the interception rate (interceptions divided by pass attempts) did. It rose from 1.66% for teams leading by 15-plus to 3.54% for teams trailing by 15-plus, gradually increasing each step along the way. Because passing attempts go up as a team’s lead decreases, the total turnover rate, fumbles plus interceptions, increases as well with its line mirroring that of interception rate. To be behind is to be at increased risk of turnovers because teams that are behind often have to pass more.

Because the total interception rate for teams that trail outpaces the pass rate, you could theorize that teams that fall behind are attempting riskier passes. Outside of turnovers, big plays are the quickest way to flip a game so maybe teams are attempting more high-risk, high-reward shots downfield when they have ground to make up. But there might also be a simpler explanation—the defense is ready for those throws.

Why do defensive coaches prize third downs? It’s a chance to get off the field and give the ball back to the offense, yes, but if a defense can put the offense behind the chains—anything longer than third-and-4 is a good starting point—it has an idea that a pass is coming more often than a run. In 2018 the average offense ran the ball 55.7% of the time on all downs but passed 57.1% of the time on third downs. Increased predictability is the defense’s advantage on third down.

Defending with a lead effectively makes the opposing offense look more like it does on third down. The ratio of run-to-pass plays is similar as are the completion and interception rates. Defend with a lead larger than a touchdown and there’s an even greater similarity. While a defense doesn’t get the get-off-the-field edge of a third down, playing with a lead does tell the defense more passes are coming, it can adjust accordingly and the passes will be less efficient as a result.

This, I suspect, is why a 7-0 lead even just a few minutes into a game matters. A team enters a game with one opponent—the team on the other sideline—but as soon as it falls behind it has three opponents—the team on the other sideline, the score and the clock. Traditionally we don’t view football this way until we’re forced to late in the game, but it’s what happens as soon as the score is something other than 0-0. The authors of The Hidden Game of Football likened the fourth quarter of games to the top of a skyscraper. “It’s what you notice first,” they wrote. “But take away the foundation of the game, or the early part of the football game, and all you have is rubble.”

If you accept that analogy, building a lead then is building the foundation for a win and I suspect turnovers and predictability on offense are giving that lead its power.
Fumbles remain dumb. They’ll happen to any team at any time. Interceptions, however, require passes to be thrown and teams that trail pass more. Those teams’ turnover rates increase accordingly. The only real way to reduce turnovers is to play less football.

Nobody wants less football, except when their team is protecting a lead. Then it’s the best thing you’ve ever seen.

This is important for Nebraska in 2019.

The 2018 Huskers trailed a lot. Nebraska’s offense ran just 30.76% of its plays with a lead, 91st nationally. Clemson led the country at 74.72% and Alabama was second at 74.03%.

UCF was in that territory two years ago. The 13-0 Peach Bowl champs led on 67.97% of their offensive snaps, third nationally. The Knights’ defense was an even bigger beneficiary of the score-early-and-often approach, leading on 83.9% of its snaps. That defense was also second in the country in total takeaways.

Central Florida’s offense that year received the opening kickoff in 10-of-13 games. It scored a touchdown on the first drive of the game seven times.

Nebraska didn’t have that luxury during Frost’s first year in Lincoln, but it wasn’t for lack of effort. The Huskers started the season by winning the coin toss against Akron and electing to receive before the start proved to be a false one. A week later, the Huskers again won the toss and took the ball. Adrian Martinez’s first drive behind center ripped right along for seven plays and 52 yards before a Greg Bell fumble killed it inside the red zone.

The Huskers didn’t score first in a game last season until facing Purdue in the fourth game. That was a loss as were two others when Nebraska struck first. The Huskers only finished 3-3 with that advantage in 2018, but their turnover margin, minus-two on the season, was plus-four in those games. That correlation probably isn’t particularly strong in this case. Strange things can happen with a small sample size. But the overall correlation between score, time and turnovers is key for any team.

The national consensus around Nebraska entering 2019 is that it will be improved. The Huskers have another year in the system, a really good quarterback and strength on the defensive line. Those are all nice advantages to have.

How much better the Huskers are, however, might be a matter of how quickly they can start letting the game of football work for them.

It might not be long at all. Frost is a fan of fast starts.

  • Never miss the latest news from Hail Varsity!

    Join our free email list by signing up below.

Share via
Copy link
Powered by Social Snap