With the NFL season finally here, I wanted to share a special secret with you. This off-season, I have spent thousands of hours researching the single best way to predict if your team is going to win games. For this example, I will use the hometown Tennessee Titans.
Consider this: In the last two years, the Titans have won quite a bit. This can be attributed to a litany of different reasons, but one of the most popular choices is because of their star running back, Derrick Henry. With that in mind, I have a stat for you: Over the past two regular seasons, the Titans have racked up 15 wins and two losses when Henry gets more than 20 rush attempts. That is absolutely incredible. However, when Henry has 20 or less rush attempts, the Titans are a pathetic 5-9.
Yes, you read that correctly. Tennessee wins 88% of their games when Henry gets over 20 carries but only gets the win in 35% of games when he doesn’t. Easy plan, right? Just pound the rock with Henry and let the 6-foot-3, 240 pound monster dominate another defense.
If you ask the vast majority of University of Tennessee students, they will completely agree. Unfortunately, they are wrong. But why? It’s not because I am lying to you. Those numbers are accurate and not intentionally misleading. Unfortunately, as many of you who have taken STAT201 know, statistics can be hard to grasp.
Sure, what I led with is technically true. However, I want you to think about something. For those games that the Titans are dominating in, are they winning games because they are rushing the ball, or are they rushing the ball because they are winning? Unfortunately, the answer to the question is the latter.
As Ben Baldwin has shown, this effect only occurs late in games, when the contest has already been decided. In the first quarter, whether your team runs the ball just once, or as many as 15 times, has very little effect on how often you win. I mean that if my team played last week and ran the ball once in the first quarter, and your team played and ran 10 times, we would both have a near identical chance that we won.
In the fourth quarter, however, things changed. Teams that had an 80% chance of winning ran the football between 10-20 times in the fourth quarter alone. On the other hand, teams with a less than 20% chance of winning ran the ball between 1-5 times in the quarter.
Intuitively, this makes sense. The team with the lead wants to call a safer type of play than also drains the game clock, resulting in less opportunity for the other team to come back. As for the team losing the game, they want to limit running the ball in order to score faster in less time.
This is the issue with Henry’s statistics. Sure, he is undoubtedly one of the best in the entire world at his position. Nevertheless, Henry is getting carries because his team is ahead, his team is not ahead because he is getting carries.
Fortunately, there are new tools that allow us to evaluate each individual play and player in order to truly maximize the odds of winning a game. For example, Expected Points Added (EPA) is a revolutionary way to determine how much impact a single play had on scoring points for a team. This is also the backbone of many advanced statistics, such as win probability models and ESPN’s Quarterback (QB) Rating. Further, Rushing Yards Over Expected (RYOE) is a brand new value that measures how much better or worse a running back performs than what the average running back would have done, given the same circumstances.
“But Max, you still need to run the ball. Henry is a total monster, and it wears out the defense. Plus, it makes play-action easier and it demoralizes the other team. That stat you started with is real, and the Titans need to make sure Henry gets over 20 carries.”
Okay, hypothetical person I just made up, I totally hear you. So let me end this with one final example.
Do you really want to win games? Do you really, really, really want to win games? Like, undefeated every year type of winning? Alright, let me tell you how. From my own personal experience, teams that have at least three Quarterback Kneels win every single game. I have seen hundreds of football games, and when a team can just get three QB kneels, they win.
So from now on, let’s just have the Titans start each game with three QB kneels! If we can just hit that magic number, we should be able to absolutely dominate.
Max Thompson is a junior majoring in business management and journalism and electronic media. He can be reached at email@example.com. Follow @The_Out_Route on Twitter for high-quality NFL analysis!
Columns and letters of The Daily Beacon are the views of the individual and do not necessarily reflect the views of the Beacon or the Beacon's editorial staff.