Analytics: Wisconsin Football & Expected Points Added (EPA)
What is EPA? I'll break it down in detail so you understand it. Then I'll dive deep into who the EPA Leaders are for Wisconsin Football
I’m back with Part Nine & Ten of my Pre-Season Series: TOP 10 Wisconsin Football Stats you need to know for the 2023 Season!
Today, we will discuss what "EPA" is and who our EPA leaders are coming into the 2023 Wisconsin Football Season!
For Part 1 | Part 2 | Part 3 | Part 4 | Part 5 | Part 6 | Part 7 | Part 8
If you follow Football Analytics, I’m certain you’ve seen the term “E-P-A” thrown around or as labels in Graphs depicting top players in the NFL & CFB!
I’m here to explain what EPA is as Part 9 of this series. And then, I will share with you who the EPA leaders are for the Wisconsin Badgers heading into the 2023 Season. I’ll also dive into some deeper analytics into what can help drive EPA for our Badgers as Part 10 of this series!!
Part 9 - What is Expected Points Added (EPA):
Expected Points Added, or its shorthand acronym EPA is a metric that defines the success/failure of each Play for each Team. It derives its value from the change in “states” resulting from a Play. In its simplest form, a “state” is the Down and Distance to reach a 1st Down and the Yards to the EndZone.
Let’s use an example state.
Example:
1st Down & 10, 75 Yards To Go (Own 25YD Line)
From this “state” and the subsequent play. Say a 4-yard Rush by the Running Back it would bring us to a new “state” 2nd & 6, 71 Yards To Go. Each of these “states” has an expectation of the points a Team will score based on historic performances in that “state”. And so, EPA is the difference in those expectations, which can be POSITIVE or NEGATIVE for each team depending on the result.
Here is a graph that will really highlight this:
Light Blue = 1st Down
Dark Blue = 2nd Down
Orange = 3rd Down
Purple = 4th Down
So you can see that losing downs and being in the same Yards to Goaline position severely deteriorates your EPA. But this graph is an aggregate of all possible states for each down. So, this graph is a broad aggregation!
Going back to our example, the actual EPA of that 4-yard rush is this:
So a 4-yard Rush, actually loses the expectation for points scored by the offensive team by about 0.11 points.
Let’s say on the next play; we also have a 4-yard rush expressed as State 2 & State 3:
Again, a 4-yard Rush, loses the expectation for points scored by the offensive team by about 0.11 points. Interesting.
Let’s say on the next play, we only have a 2-yard rush, but pick up the 1st down? How will that impact the EPA?
Well, you can see that play is worth about 1.05 in EPA! But it was 2 yards less than the other two plays. Yes, but by moving the sticks and earning their team another 4 downs to continue moving the ball, that play in a vacuum was worth so much more!
Potential Issues with EPA:
But what if each play had a different ball carrier? Why would the 1st two RBs be downgraded for gaining 2 More Yards/Attempt? This represents a major potential pitfall to using EPA as a standalone metric to evaluate players because such context matters immensely, as I’ve just shown.
But if it were all one RB, say Braelon Allen. That would be 0.83 EPA over those three carries and would be about 0.28 EPA/Play.
So, over the course of a Game, Season, or Career, you would assume that these potential deviations from normalcy I presented wouldn’t have such an outsized impact. Thus, you could use the EPA metric as a way to value a Player’s Total Impact & Impact/Play.
Part 10 EPA Leaders - Wisconsin Football QB Tanner Mordecai’s EPA:
We see Tanner Mordecai is really good with 0.12 EPA/Play over his 2021-2022 Seasons at SMU. And while a majority of those games are really positive, we can see two games ironically against Cincinnati where he really struggled!
What drove his EPA/Play so low? Well, I tried to correlate some factors to it!!!
INT Rate & EPA/Play:
Not too strong a correlation here. While there is definitely an impact on EPA by throwing INTs, there is still more to it!
TD Rate & EPA/Play:
There is a much stronger correlation here, but is there a way to get it even tighter? I think so!
Net Big Play Rate & EPA/Play: (Net Big Play = [TD Rate - INT Rate])
So, it's not shocking stuff here, but increasing your TD Rate and limiting your INT Rate will skyrocket your EPA/Play as a QB!
What does that look like if Mordecai sits in the large part of this distribution and we project this correlation? It'll look something like this:
I’d say we’d be good with any of those stat lines?? But I called out his Heisman chances in a prior post.
How do these numbers stack up?
On a 32 ATT/G Average (Longo’s Average at UNC), if Mordecai were to achieve >11% TD Rate & <3% INT Rate, he could be in that conversation. Although that would put him on the low end of the spectrum (Bryce Young 2021). But it would be Heisman Level output!
However, does 49 TDs seem very achievable? I’m going with no. But if he can keep his INT Rate below 1.75%, getting 35+ TDs would get him in that conversation!
Wisconsin Football RB’s EPA:
Keep reading with a 7-day free trial
Subscribe to Badgernotes to keep reading this post and get 7 days of free access to the full post archives.