I don’t blame anyone for believing the things they do about football.
Football is emotional. We invest much of ourselves rooting for our favorite teams — deeply hoping that at the end of each game, they will be victorious. And when they’re not (like the Kansas City Chiefs against the Indianapolis Colts on Sunday night), we become distraught.
As humans, it’s just our nature that when things don’t go our way (and sometimes even when they do), we search for patterns — something on which we can blame the tragedy. We’re really good at it.
So good, in fact, that we often find patterns that don’t actually exist. And one of the mythical patterns humans find after their NFL team loses a game is time of possession.
The argument generally goes like this:
One team successfully ran the ball (or the other team failed to stop the run), allowing them to take time off the clock with each possession. This prevented the other team’s star player (or players) from having the ball. The star player is very good at football, so limiting their time with the ball is a good thing. At the end of the game, the team employing this strategy not only had more points, but also far more minutes holding the ball. This was the reason they won the game.
Allow me to repeat: I do not blame (or judge) anyone for believing this argument; you hear it all the time. But here’s the thing: believing something doesn’t make it true.
Let me show you what I mean.
There are 60 minutes in a football game and two teams, so it’s zero-sum — and thus, it is true: one team having the ball longer does mean the other team has the ball less.
But what effect does that have?
The Colts had a possession in which eight minutes and 34 seconds came off the clock. They ended this drive by kicking the field goal — and then they kicked the ball back to the Chiefs.
And that’s what what makes football fun: after each team’s possession (or turn) with the ball, the other team gets it back.
After that long drive, what was it about the Colts’ run offense — or the Chiefs’ run defense — that prevented the Chiefs offense from having their own drive that consumed just as much time? Why couldn’t the Chiefs just run the ball constantly — maybe even throwing in a few quarterback kneels here and there — to take as much time off the clock, therefore evening out the time to possession?
In fact, the Chiefs have repeatedly shown that they don’t care how long they have the ball on offense; against the Oakland Raiders, they scored four times in a matter of minutes. They take more deep passing shots than almost any team in the league — and that’s the least effective way to take time off the clock.
Why? Because the goal of the offense is to score points — not take time off the clock. It is the team with more points that wins — not the team that has had more time with the ball.
You might argue it is not actual time of possession that matters here, but rather that the Colts running all over the Chiefs defense limited the Chiefs’ number of possessions — therefore giving the Chiefs fewer opportunities to score.
But don’t forget what makes football fun: after each team’s possession (or turn) with the ball, the other team gets it back. No matter what kind (or how many) plays they run, both teams get an equal number of chances on offense.
There are exceptions to this, of course. A turnover on a punt or kickoff gives a team an extra possession. And it’s possible for a team to have more possessions than the other simply because they started (and ended) a half with the ball.
But that didn’t happen on Sunday. According to the NFL game summary, both teams had 10 turns with the ball. Here’s what the Chiefs did with theirs:
- Field Goal. (3 points)
- Touchdown (7 points)
- Quarterback kneel — end of half
- Turnover on downs
- Field Goal (3 points)
The Chiefs averaged 1.3 points on their drives. The Colts averaged 1.9. So we can see that by scoring more points per drive, the Colts ended the game with 19 points to the Chiefs’ 13.
In fact, we could go back to every football game ever played, and I’d be willing to bet that in only a tiny percentage of them has the team with more points per drive failed to win the game. The only times it’s likely to have happened are when a team scored defensive or special teams touchdowns — which count as points but not as possessions.
Don’t believe me? Here’s the drive data for the NFL last season.
Offensive Drives - 2018
There wasn’t a single team that averaged even one drive per game more (or fewer) than their opponents. The Chicago Bears had the greatest disparity with 13 additional drives — but that’s still less than one additional drive per game; most teams have less than five more (or fewer) drives than their opposition.
Also notice how the New Orleans Saints— the league’s best at points per drive scored minus points per drive allowed in 2018 — are tied with the Los Angeles Rams for the league’s best record. With some exceptions owing to the factors I mentioned before, the record rankings generally follow the differential in scoring per drive — the teams with the best records have the greatest differentials, while those with the worst records have the lesser ones.
Now — there is one argument floating around about time of possession that actually has some statistical merit: that limiting the total number of possessions (the number of turns both teams get with the ball) can be a good thing for an underdog team — like the Colts on Sunday night. So for the sake of fairness, I’ll present this argument. But then I’ll explain why it doesn’t apply in the NFL — or at the very least, why it didn’t apply in Sunday’s game.
Imagine you are in a three-point shooting contest with Steph Curry — the NBA player known for his ridiculous accuracy beyond the arc. You aren’t a better three-point shooter than Curry; both of you know this. Curry’s career three-point percentage is 43.6%. Let’s be generous and propose yours is about half that — say 20%. But to make the three-point contest a bit more interesting, you and Curry have agreed that you get to decide the total number of shots each of you take. If you choose 100, you each shoot 100 shots. If you choose 50, you each shoot 50. The winner is whichever shooter makes the most shots.
What number of shots would give you the best chance of winning?
This is an example of what statisticians call a Bernoulli Trial, which is when you have a fixed number of experiments (three-point shots) each with two possible outcomes (a make or a miss) and an equal probability of success (a make) on each one.
The reasoning goes like this: if given 100 shots, Steph Curry is going to beat you. We can actually calculate the exact percentage of this happening with a simple line in a spreadsheet like Excel. If you give Steph Curry 100 shots, there’s a 99.999% chance he will beat you. However, if you both get just one shot each, you have an 11% chance of winning the contest; it’s the probability of him missing the one shot (1 minus 43.6% — or 56.4%) multiplied by the probability of you making one shot (20%). That’s 11.3%.
So some will argue that in a situation like Sunday’s game, the Colts were simply trying to limit the total number of possessions. The Chiefs would be like Steph Curry — an offense more likely to score points. The Colts would be like you — less likely to score with each opportunity. You decide to force fewer total chances, which (you think) means you have a better chance to win.
And in a teeny-tiny way, you do.
Here’s the problem: NFL teams typically get 9-13 possessions per game; we aren’t dealing with the difference between 100 shots and just one. Through four weeks, the Chiefs averaged 10.25 drives per game, so reducing that number by 0.25 just doesn’t make a significant difference.
Besides... Patrick Mahomes specifically only had eight possessions against the Jacksonville Jaguars (as Matt Moore came in for two drives), yet still put up 37 points. How many did Mahomes have against the Colts? 10. So the Colts didn’t actually limit the Chiefs’ total possession numbers any more than any other team this year.
While it’s a fun thought experiment, the Steph Curry analogy can’t be argued to apply here. Coming in to this game, the Chiefs offense averaged 3.29 points per possession. Sunday night — as we noted before — it averaged only 1.3. Looking at those numbers, is it really fair to say the pivotal battle was the one between the Colts rushing attack and the Chiefs rushing defense?
And let’s be honest — if Curry had a hurt ankle and Cameron Erving was responsible for protecting him, maybe we would want him taking 100 shots!
I’ve seen others argue by increasing your time of possession, you can wear down the opposing defense — thereby making them less effective. But this has been emphatically proven to be false. Plain and simple — despite what the announcers say on TV during the fourth quarter — defensive production is not impacted by rest time.
There is, however, obvious value in being able to run the ball well when your team is winning in the fourth quarter — as the Colts were. In this case, a better defensive performance by the Chiefs may have gotten the ball back with enough time for the offense to put some more points on the board and tie the game.
But if the Chiefs offense had been able to score points on a single one of its six drives in the second and third quarters, Indianapolis would not have had the ability to run the clock out; instead, they would have had to respond by scoring points themselves.
So... no. The Chiefs didn’t lose to the Colts simply because the Colts dominated time of possession by running the ball against a bad Chiefs run defense.
That isn’t to say that the Chiefs run defense isn’t bad. No matter how advanced a metric or statistical technique you use, the Chiefs run defense is bad — and if it had been better against the Colts, they would have had a better chance to win.
But on Sunday night, the defense as a whole did hold the Colts below their average points per drive for the season. Unfortunately, the Colts defense did even better, holding the Chiefs offense to their lowest points per possession of the Mahomes era — and that is what made the difference in the game.