An Explanatory Analysis

Welcome to part two of my deep dive into Baker Mayfield’s young (but fascinating) NFL career! In part one I explored the fledgling QB’s statistics through two seasons and observed three distinct periods – early struggles under the Hue Jackson regime, an ascendant second half of 2018 with Freddie Kitchens calling the plays, and a sophomore slump that ultimately lead to Freddie’s firing.

In this installment, I continue down the rabbit hole in an attempt to understand why – why Baker stunk it up with Todd Haley as OC, why his rookie campaign turned around under Kitchens, and why the Browns passing offense stagnated in 2019. Once the why’s are covered, I look to the future by examining the pairing of Baker and his newest play-caller, former Vikings OC Kevin Stefanski.

2018: Weeks 3-8

Mayfield kicked off his NFL career with a bang, leading the Browns to a comeback victory over the Jets in week three after replacing Tyrod Taylor following a concussion. However, if you read part one or remotely follow the NFL you will know that the good times would not last for long. Baker and the Browns offense regressed in the following weeks, losing four of their next five games before Jackson and Haley were ultimately shown the door.

What went wrong with the Browns offense? Why were drives stalling? Why was Baker taking a beating seemingly every time he dropped back to throw? Fortunately, the football-viewing public now has access to granular team and player-level data which helps us begin to answer some of these questions by examining the relationship between play-calling tendencies and offensive production. Because this article is focused on Baker Mayfield, I will limit this investigation to passing metrics with data courtesy of Sports Info Solutions (SIS) and nflscrapR, including the following variables:

• play-action rate
• shotgun rate
• screen rate
• bootleg rate
• dropback depth
• average depth of target (aDOT)
• average number of wide receivers

The following table shows the 2018 Browns passing offense from weeks 3-8 (the weeks Baker played under Jackson and Haley) with accompanying rankings relative to the other 2018 NFL passing offenses.

After exploring the splits for each of these variables, something stuck out about Baker’s performance on plays with deep dropbacks, specifically 5-7 step drops out of shotgun. Through week 8, Mayfield scraped together a passer rating of 79.9 with a 40.8% success rate (plays resulting in positive EPA) on his 240 total dropbacks. However, if we just look at his 22 deep dropbacks from shotgun his passer rating plummets to 30.4 and success rate drops to 22.7%. Although these plays constituted less than 10% of Mayfield’s total dropbacks, they had a sizeable impact on Baker’s aggregate stats, accounting for -10.2 of his -30.3 total expected points added (EPA) through week 8.

It should also be noted that Brett Kollmann found a similar finding in his film study which inspired this whole writeup, The Ballad of Baker Mayfield. Once again, I highly recommend checking out his content.

Combining Kollmann’s analysis with a bit of my own film review, I determined that each failed deep dropback contained one or more of these three elements:

• Subpar pass blocking, especially on the edge
• Lack of pocket presence from Baker
• Poor timing between Mayfield and his receivers

Let’s take a look at some examples! Here is one such play which came on a big third down late in the game against the Ravens in week 5.

This play was doomed from the start. Mayfield is barely able to complete his dropback before his own left tackle, Desmond Harrison, is in his lap after being dominated by a bull rush from Terrell Suggs. Although Baker is able to improvise and get the completion, the Browns are well short of the sticks and the drive stalls.

Here’s another deep dropback on third down that ended in disaster, this time against the Bucs in week 7.

Although the pocket collapses quickly, Mayfield is not without fault on this one. If Baker keeps his eyes downfield, he would have seen a wide-open receiver streaking down the middle of the field just as he reaches the top of his dropback. Instead, Baker succumbs to the pressure and once again, the drive stalls.

And for our last ugly example - here is a failed third down vs. the Chargers in week 6.

The Browns actually have decent protection on this play. Unfortunately, by the time Antonio Callaway is able to regain his footing after stemming his drag into a go route following the mesh point, Baker is unable to see him and get the ball out, instead going down for the sack and ending the drive.

Part of the blame should go to OC Todd Haley on this one. This long-developing play probably resulted in many long receptions from Big Ben to AB when Haley was with the Steelers. However, Haley needed to recognize that Antonio Callaway is no Antonio Brown and Baker’s ability to step up into the pocket (both due to his height and inexperience) is nowhere near that of veteran Ben Roethlisberger.

At a certain point Browns management couldn’t stomach anymore of this garbage and finally kicked Hue Jackson to the curb along with Todd Haley to be replaced by Gregg Williams and Freddie Kitchens as interim HC and OC, respectively.

2018: Weeks 9-17

The turnaround wasn’t immediate, but it did not take long to notice that Baker and the Browns offense had turned a corner with Kitchens calling the plays. As we saw in part one, Baker went from the bottom of the statistical charts under Haley, to one of the league’s top performers under Kitchens. What was Freddie doing with his play-calling to bring out the best in Mr. Mayfield?

The table below shows the Browns’ passing play-calling tendencies for the second half of the season and how these changed with Freddie calling the plays.

The two changes that immediately jump out are the increase in play-action usage and decrease in deep dropbacks. The Browns went from one of the lowest play-action rates during Baker’s first few starts to surpass the league average under Kitchens. Play-action has received increased attention in recent years due to its strong correlation with offensive efficiency, leading some to argue PA is underused at the NFL level. Therefore, it makes sense to assume that this played a part in the turnaround. Interestingly, the data does not seem to back up this idea as Baker’s splits were largely similar with and without play-action.

Once more, deep dropbacks seemed to be a pivot point. Although the Browns’ struggles on deep drops from shotgun continued into the second half of 2018, Kitchens mitigated this problem by simply calling fewer of these plays – instead, opting for shorter drops and more screen passes. Below is one of these shorter dropbacks from the Browns’ thrashing of the Falcons in week 10.

These quick hitters not only took pressure off the pass blocking up front but also helped to simplify Mayfield’s reads. In this case, Baker is watching the outside backer who takes the TE in the flat. Once he sees the backer commit, Baker zips it right behind the LB’s head, hitting the slant route for a big gain.

Although Kitchens shortened up Baker’s dropbacks, the Browns were still able to stretch the field vertically as the team’s average depth of target actually increased in the second half of 2018. The play below is an example of the Browns taking shots downfield while also getting the ball out quickly.

The Browns’ improved pass timing shows up on this Mills concept vs the Ravens in week 17. Previously, receivers were still trying to get open as Baker was reaching the top of his dropback, throwing off the rhythm of the play and often leading to negative outcomes. However, here we see receivers making their cuts just as Baker plants his back foot. Mayfield spots the safety bite on the crossing route, takes one quick gather step, and hits Breshad Perriman on time and in stride for the long TD.

The Browns offense was flying high to finish 2018. So much so that the team opted to roll with Freddie Kitchens again in 2019 – not just as offensive play-caller, but also head coach. Jimmy Haslam and company went with the “if it ain’t broke don’t fix it” approach. What could possibly go wrong?

2019

Expectations couldn’t have been higher for Baker and the Browns going into 2019. The strong finish to 2018 combined with the acquisition of star wideout Odell Beckham had fans dreaming Superbowl. Sadly, once the season started, it didn’t take long to recognize that these playoff fantasies were actually pipe dreams.

In order to begin diagnosing the Browns’ reemergent offensive woes, let’s turn to our handy dandy passing play-calling tendency table.

We’ll start by looking at play-action. Much like the season before, Kitchens continued to lean on play-action in 2019 – and for good reason. Mayfield’s 2019 passer rating when faking a handoff (roughly 29% of dropbacks) was more than 30 points higher than on plays without play-action. Although the Browns were above the league average in terms of play-action usage, this massive difference in Baker’s splits suggests Freddie could have used even more fake handoffs than he did in 2019.

We also see that the percentage of deep dropbacks reversed trend and moved back upwards in 2019. Unsurprisingly, Baker and the Browns offense continued to struggle on these plays just as they did throughout 2018. While Baker’s aggregate stats weren’t great in 2019 (80.1 passer rating, 42.7% success rate), his numbers on deep dropbacks were downright bad (60.5 passer rating, 34.5% success rate). The same structural problems were still there for the Browns – poor exterior pass blocking, an antsy Mayfield in the pocket, bad receiver timing. Yet, for some reason, after watching these inefficient, long-developing plays consistently fail over the previous two seasons, Freddie continued to call them.

Another interesting trend in the Browns’ 2019 play-calling was the shift away from intermediate dropbacks in favor of shorter drops and run-pass option plays (RPOs). On the face of it, this appears to have been a prudent decision by Kitchens as Baker’s efficiency stats on these short drops were noticeably improved compared to intermediate dropbacks. However, this strategy is not without its downsides. Although shorter dropbacks often lead to increased completion rates, focusing too much on quick hitters can limit an offense’s ability to stretch the field vertically, which is what we observe when we look at the Browns’ average depth of target (aDOT) which went from one of the highest in 2018 to just above average in 2019.

Of all his 2019 play-calling sins, unoriginality was perhaps Freddie’s greatest transgression. If we compare Kitchens 2019 play-calling tendencies to the year before, most metrics were either unchanged or continuations of the trends we saw in the second half of 2018 (more play-action, more screens and quick throws).

Kitchens’s play-calling was also unremarkable when we compare it to the other passing offenses around the NFL. Below, I have constructed a principal component analysis (PCA) much like I did in part one, but this time we are observing 2019 play-calling tendencies in the passing game rather than aggregate QB passing stats. Again, PCA is a statistical method that “compresses” correlated multidimensional data by creating new uncorrelated variables or principal components (PCs) which capture the most variance in the data. The first principal component, PC1 corresponds to the line drawn through our 7-dimensional vector space which explains the most variance (i.e. best fit line, explaining 29% of the variance in this case) with each subsequent PC corresponding to the perpendicular axis explaining the next most variance. I know that this is a lot to digest, but for now let’s just see what this looks like when we examine the first two principal components.

In essence, teams further from the origin are considered more unique in their pass play-calling while teams closer to the origin are less unique or basic, as some would say. Interestingly, the Browns are nearly right on top of the origin, as basic as you can get, which isn’t exactly where you want to be in the modern NFL. If we compare offenses near the origin (Browns, Colts, Jets, Lions) to those furthest from the origin (Vikings, 49ers, Saints, Ravens) there would appear to be some connection between uniqueness and offensive success. However, before making any definitive conclusions, I must note that this scheme analysis is far from complete as it is missing many crucial elements that define an offense like pre-snap motion or route combo usage, just to name a few.

Luckily, we will get to see this “uniqueness leads to success” idea be put to the test with the Browns offense in 2020. After cutting ties with Kitchens in January after just one season, the Browns opted to replace Freddie with Kevin Stefanski, a key architect behind the Vikings’ distinctive passing offense in 2019.

2020+

This coming season will likely be a make-or-break campaign for Mr. Mayfield. If Baker stinks it up for the second consecutive season, the Browns are liable to blow it up once more and move on from the third year QB entirely – especially with a chance to grab Justin Fields or Trevor Lawrence in the 2021 draft. On the other hand, if the pairing with Stefanski is a success and Mayfield is able to rekindle some of that 2018 magic, the young gunslinger could earn himself a nice payday in the form of a long-term extension next offseason.

With this in mind going into 2020, everyone around Cleveland wants to know if Baker can gel with Stefanski’s unique passing scheme. To begin answering this question, let’s start by taking a look at the Vikings’ passing tendencies for 2019.

I’ll begin with play-action, as is tradition. Although the Kitchens-lead Browns leaned heavily on PA in 2019, Stefanski and the Vikings still managed to outpace them, attempting a fake handoff on 30% of all dropbacks. This trend bodes well for Mayfield if it continues into 2020. As I showed earlier, the young quarterback had significantly better numbers in 2019 when using PA compared to straight dropbacks.

Another defining characteristic of the Vikings 2019 offense was Stefanski’s clear preference to have his QB take snaps from under center rather than in shotgun. This aversion to the gun resulted in a league-low 50.7% shotgun rate on passing plays. To put this figure into perspective, the next lowest shotgun rate was 65.2%, belonging to the 49ers. At first blush, this would appear to be a big issue for Baker due to his history in shotgun-heavy offenses in college and high school – a fact often listed as a potential concern during the draft process. Surprisingly, Baker has actually been worse from the shotgun during his brief NFL career, with a passer rating nearly 25 points lower than when passing from under center. Barring a major regression in this aspect of his game, it looks like Baker shouldn’t have any problems taking a few additional snaps under center in 2020.

Let’s return one last time to the topic of deep dropbacks, of which the Vikings used more than the Browns in 2019. Again, knowing what we know about Baker and his poor play on deep drops, this would not appear to work in his favor. However, not all deep dropbacks are created equal, which is why I was careful to mention that Mayfield’s struggles on deep drops were largely limited to shotgun snaps. Baker actually has surprisingly good stats on deep dropbacks when under center (125.2 passer rating). Knowing this information, the outlook brightens when we just examine dropback tendencies when in shotgun. Removing under center snaps causes the Vikings to drop below the Browns to the ninth lowest deep dropback rate in the league for 2019.

The last interesting play-calling tendency we will dig into is Stefanski’s heavy bootleg usage. Stefanski’s entire offense is built around the outside zone and when this play is clicking, defenses are forced to sell out to stop the run. This is when Stefanski strikes by sprinkling in play-action bootlegs to devastating effect – as evident by Kirk Cousins 123.5 passer rating on these plays in 2019. But what does this mean for Baker? To answer that we need to look into Baker’s track record when operating out of the boot. Although the sample size isn’t huge, Browns fans should be happy to know that Mayfield was a solid 12-12 on bootlegs last season with a perfect passer rating.

As mentioned earlier, because he is on the shorter side, Baker struggles to see the field at times from within the pocket. However, this problem is almost completely mitigated when Baker is out on the edge running the boot which allows him to scan the field without interference.

Below is one such example of a Stefanski-style bootleg which Baker ran to perfection against the Bills in week 10 of last season. After showing this split zone look earlier in the game, Baker keeps this one on a naked boot and hits Jarvis Landry for a big gain after the latter slipped his block on the would-be backside DE.

Hopefully, if all goes to plan, we will see plenty of plays just like this once Stefanski gets his hands on this loaded Browns offense in 2020.

Conclusion

Mayfield and the Cleveland offense never managed to get going under Todd Haley largely due to too few play-action fakes and too many deep dropbacks out of shotgun – a play which Baker clearly seems to struggle with. Once Haley was replaced with Freddie Kitchens, the Browns began getting the ball out quicker which resulted in fewer sacks and more on-time throws to open receivers. Unfortunately, Freddie turned out to be a one trick pony and was unable to adjust his play-calling in 2019 to bring out the best in Baker and the Browns offense resulting in his firing to finish the season. Now, in steps Kevin Stefanski with his distinctive offensive scheme, heavily reliant on bootleg play-action, especially from under center – some of Baker’s most successful plays at the NFL level.

2020 looks to be a decisive year for the fiery young quarterback. If Mayfield can master this new offense and regain some of his late-2018 form, the franchise-QB-starved Browns may just throw the checkbook at him next offseason. Alternatively, if Baker fails in this new scheme seemingly tailored to his skillset, 2020 may be the beginning of the end for the mysterious Mr. Mayfield.