Let’s talk about ESPN’s Football Power Index or “FPI” as many call it.
Yes, you’ve probably heard people reference it 100 times already this season, but for sports fans who are interested in how it works, you’ve come to the right place.
When you go to your team’s page on ESPN and click on the next upcoming game, you’ll see the Matchup Predictor — something you can’t miss. Here it gives a percentage of how confident the model thinks your team will win the game. The higher the percentage, the better.
No matter which matchup you are looking at, the worldwide leader in sports has you covered, with a forecast for every game that hasn’t been played yet. Now that you’ve come this far, let’s dive in to see how it works for college football.
Week 5 Bowl Projections: Potential postseason destinations for BYU, Utah, & USU
What Is The Football Power Index?
In 2013, ESPN unveiled its Football Power Index system. It’s important to note and remind you that FPI is not a poll or ranking system, it’s a predictive rating system crafted to assess team prowess and forecast their performance moving forward.
Unlike conventional rankings that aim to simply stack teams from 1 to 130+, the primary objective of FPI lies in its ability to make accurate predictions about game results and the outcomes of entire seasons. To put it in another perspective, if sportsbooks ever published power rankings, it would resemble FPI in many aspects.
What Goes Into The Football Power Index?
The Football Power Index uses a sophisticated rating system that breaks down each team’s overall strength into three key components: offensive, defensive, and special teams. These components are designed to estimate the expected contribution of each unit to the team’s net scoring margin, assuming the game is played on a neutral field against an average FBS (Football Bowl Subdivision) opponent.
During the preseason, FPI relies exclusively on historical data from previous seasons to form these components. This data includes factors such as the number of returning starters, past team performance, coaching tenure, recruiting rankings, etc. These preseason calculations enable FPI to offer predictions and evaluate the strength of a team’s opponents right from the start of the season. Of course, as the season progresses and based on how teams play (wins, point differential, etc.), the weight assigned to preseason data diminishes.
It’s worth emphasizing that while preseason data becomes less influential over time, it is never completely gone. This is because historical information continues to contribute to the accuracy of FPI’s predictions, even as the season unfolds. This practice aligns with how Vegas oddsmakers also factor in prior performance when setting their lines, underscoring the significance of incorporating past seasons’ insights into assessing team strength and performance.
Big 12 Football Power Rankings: BYU & West Virginia moving up, Kansas schools fall
Football Power Index Factors For Preseason Ratings
Preseason polls and rankings continue to get worse with each passing year. For example, last season, of the 25 teams that were ranked in the AP Preseason poll, only 10 of those schools (Georgia, Michigan, Alabama, Ohio State, Clemson, Pitt, Notre Dame, Oregon, USC, and Utah) finished the season ranked. And the trend continues to get worse with each and every year.
However, as you’ll see later, ESPN’s Football Power Index has an accuracy rate that is surprisingly better than what you probably think. Before we get to that, here’s a look at how the FPI comes up with its preseason ratings for teams. There are four components to the preseason rating: past performance, returning starters, coaching tenure, and recruiting rankings. Here’s a breakdown of each:
Past Performance — The most recent year’s performance is by far the most important piece of information powering preseason FPI. Having said that, the system still looks at the past 4 years but gives the most weight to the prior season.
Returning Starters — This includes everyone: Offense, defense, and special teams. But not all starters are created equal. There is special consideration given to starting quarterbacks or transfer quarterbacks with starting experience. This is the second piece of information powering preseason FPI. Because starters interact with other inputs, it’s not as simple as saying an extra returning starter is worth one point. Nonetheless, a starting quarterback is worth about 3.3 points per game to a team returning an average offense (all else equal), and a transfer quarterback is given half the weight of a starter.
Recruiting Rankings — Did your team land that 5-star recruit? If so, it got a bump in the recruiting rankings as FPI looks at four recruiting services — ESPN, Rivals, Scouts, and Phil Steele — to measure the talent on a team’s roster.
Coaching Tenure — Is your team led by a longtime coach? If so, it got a boost according to FPI. If not, the rankings in each phase (offense, defense, and special team) will take a drop with a new or inexperienced coach.
Notable: There are also many other data points that go into FPI’s projections. For example, an additional week of rest (bye week) gives teams a boost while every additional 1,000 miles traveled for road games decreases the chance of a win. Each team’s schedule is simulated 10,000 times in order to get an accurate assumption of what will most likely play out for the season.
How Accurate is FPI?
Like most rating and predictive systems, the Football Power Index is not perfect, however, it does a pretty good job nonetheless. Over the past decade, the team that the FPI favors has gone on to win around 74% of the time. This is also pretty comparable to the results from sportsbooks. If you want to follow along with how FPI performs throughout the season or check out other predictive systems, feel free to go to the prediction tracker website. Here you see the current stats for the year, as well as previous years.
Of course, no system will be 100% accurate — if you find one let me know — and every year there will be teams that surprise and teams that disappoint. A great example of this last year was TCU making the national title game. The Horned Frogs weren’t ranked in the Top 25 and FPI had TCU going 6-6, with a 72% chance to make a bowl. Even Big 12 member Kansas State was projected to go 6-6 last year and they ended up playing TCU in the Big 12 Championship game.
While it can certainly whiff on calls like that, overall, its 70%+ accuracy rate has shown that it can still be useful and isn’t as bad as some people make it out to be.
