Carson Wentz Was Who He Was . . . Or Was He?

Thanks to the nflFastR project and NFL NextGen Stats for the timely sources of data.

---

I’m sure you have heard the phrase “Past performance is not necessarily indicative of future results”. While that statement is almost universally true, there is a corollary that people often ignore, that is just as true: “Past performance is frequently indicative of future results”.

After the 2021 trade that brought Carson Wentz to Indianapolis, I had expressed concerns about his ability as a QB, based on his past performance. At the time, I wrote:

He’s got a lot to prove to me . . . In his 5 years in the league, Wentz’s EPA efficiency has ranked 20th, 2nd, 15th, 18th, 31st. — Stampede Blue Writers Roundtable: Carson Wentz Trade

Those rankings show a QB that has had a lot of inconsistency in his game and is overall, just average. So, it shouldn’t have been surprising to anyone that in 2021, Wentz had a lot of inconsistency in his game and was overall, just average.

---

Top Line Comparison

The Expected Points Added (EPA) of each pass play can be used to create 2 very powerful QB measures:

1. EPA per drop-back (the average value of pass plays)
2. Passing Success Rate (the % of pass plays that have positive value)

To give a better understanding of what are good and bad measures of these metrics, I took the top 32 QBs by drop-back volume from each year since Wentz entered the league (2016-2021) and calculated their cumulative yearly average PSR and EPA/d along with the yearly avg minimums and maximums. Here are those numbers, along with Wentz’s 2016-2020 total numbers, showing where he fell in those ranges while he was with Philly.

Wentz split the ranges almost exactly down the middle. Using the min/max values of PSR, the mid-point is 46.0% and Wentz eclipsed that by a mere 0.1%, which was 1.1% below the mean. The mid-point for EPA/d is 0.076 and Wentz was only 0.001 below that, finishing -0.027 below the mean.

In other words, prior to coming to Indy, Wentz’s career average was . . . average. So, how did he do in 2021?

He was average. Dun-Dun-DUUUUN!!

There was almost no change in either his PSR or EPA/d between Philadelphia and Indy, even though he had different coaches, different wide receivers, different O-lines, different opponents, different weather, etc. There are 2 lessons you should learn from that:

1. A QB's performance is more about the QB than the people around him
2. QBs don’t magically become someone else because they change zip codes (sit down Tannehill, no one cares)

---

Comparison Breakout

Yards is the common language of football, but there are significant game events that are not captured by yards. Touchdowns, turnovers, and first downs all have dramatic game impacts beyond the yards assigned to those plays. EPA correctly captures the actual value of those events, which is why I rely on it so heavily. However, when I roll EPA up into a single number, I lose the ability to see those individual drivers. So, let me break them out.

The following table deconstructs Wentz’s Philly EPA/d numbers into the portions that are TDs, turnovers, first downs(1), and everything else (“Other”). All the EPA values are per drop-back measures, so if you add up all the individual components for Wentz, you get his overall EPA/d. The “wgt” column is the relative weight of each event as a % of drop-backs (e.g. the league TD rate is 4.3%).

Looking at the last column, Wentz’s EPA/d on the Eagles was -0.027 below league average, and the largest driver of that was first downs, where he was 3 points below league average. The second largest driver was turnovers, which countered his poor first down value a bit by being 0.017 better than league average. These aren’t huge deviations from the mean when compared to the min/max ranges, but they show his relative strengths and weaknesses.

The “Other” category is pretty much league average, but it comprises about 63% of all NFL plays, so opening that up to more detail might be revealing. If I remove TDs, turnovers, and first down plays, the remaining pass plays in the “Other” category can be described as sacks, scrambles, incompletions, and completions(2):

The last column shows that Wentz’s EPA from sacks and scrambles is exactly league average, while his value from non-first-down completions are just a hair above average. It is only his incompletions that have negative value.

So, to summarize his Philly career: not good at getting first downs, pretty good on turnovers, too many incompletions, and everything else is meh. Sound familiar?

---

Comparison To Indy

Here is how those categories compared to his tour in Indy.

The first 2 columns are his actual EPA values and across the board. The third column is the difference in those numbers and shows that there was very little change in any of the categories.

The last 2 columns compare his numbers on each team to the NFL average, and I have highlighted the 3 biggest drivers on each team. It shows that while on the Eagles his overall passing value was:

• a bit below league average
• driven primarily by first downs and incompletions
• but offset somewhat by minimizing turnovers.

Whereas on the Colts, his overall passing value was:

• a bit below league average
• driven primarily by first downs and incompletions
• but offset somewhat by minimizing turnovers.

So . . . yeah.

---

Summary

The point is that QBs don’t deviate much from their career numbers. Prior to the season, a lot of Colts’ fans clung to the hope that Wentz would bring his 2017 game to Indianapolis, but, unsurprisingly, what we got was his 2016 - 2020 average game.

Many will point to a thin receiving corps, limited pre-season snaps, ankle injuries, and potentially COVID from de-railing his 2021. I’m not going to sit here and claim 100% that they are wrong . . . but they are wrong . . . 100%.

It’s not that those variables don’t matter, it’s just that they aren’t the only variables at play. Ask yourself this:

• If it was just poor receivers, then why did old man, noodle-arm, statue Philip Rivers have the #7 EPA/d with the same receivers?
• If it was just getting more snap volume, then why after week 5 did Wentz clearly regress as he played more snaps?
• If it was only his ankles limiting him against the Titans, then why did he put up similar poor numbers against the same team 5 weeks later?
• If it was COVID . . . you know what, I’m not getting into that one.

There are so many variables in football that it is easy to cherry-pick your way into thinking and excusing anything (even that Jacoby Brissett would have led us to the playoffs if he hadn’t been injured), but that way lies madness.

At some point, you have to stop reaching into the infinite bag of “what-ifs” and “yeah-buts” and simply believe what the numbers have been telling you all along: Wentz had almost no chance of becoming a top tier QB in Indianapolis, because his career numbers in Philadelphia showed us he was not a top tier QB. No deeper analysis needed.

I don’t know who will be under center next year for the Colts, but unless they are a rookie, my best guess is that his 2022 will look a lot like his career average numbers. I also bet a lot of people will be surprised if that happens.

---

Footnotes

1 ) TDs are officially scored as first downs. I have removed TDs from first down counts so that they are mutually exclusive.

2 ) All metrics are mutually exclusive and completely exhaustive. So “completions” and “scrambles” excludes plays that result in first downs, TDs or turnovers and incompletions excludes INTs.