Black and Gold's Consistent Fold

Written by: Eric Li (https://medium.com/@eli1)

The Pittsburgh Steelers have been touted as one of the most talented offensive teams in the entire NFL. With the top ranked wide receiver and running back in the NFL alongside a grizzled veteran quarterback, it’s no wonder that this team is labeled as a contender every year. However, ever since the inception of the Killer B’s, the Steelers have failed to even make the AFC championship outside of the 2015 season in which they were annihilated by the New England Patriots. This begs the question: are the Steelers truly an elite team, or are they just a group of overrated individuals?

Understanding the Stats

In addition to taking a look at the Steelers’ post-season performances, we will also be analyzing two advanced metrics, DYAR and DVOA, that have become more widely used in rating NFL teams and individuals. Below is a quick summary of the two statistics, paraphrased from the Football Outsiders methods.

Both statistics were created in an attempt to more accurately describe the value of plays made during games. For example, a 1st down conversion during a 3rd and 8 situation is vastly more impressive than a touchdown during 2nd and 1. The current system of evaluating player performance tends to overweigh the value of touchdowns and total yards by the offense, and fails to account for play-by-play context. In order to address this, each NFL play is given a success value. Success during any given drive can be generalized as 45% of needed yards on first down, 60% on second down, and 100% on third or fourth down. A successful play will be given one point, with fractional scores being allowed (for example 0.54 points for 8 yards on a third and 10). This is just one example of success. These success values are then added together, adjusted for the opposing defense, then normalized to give us defense-adjusted value over average, or DVOA. By normalizing the value, good offense is more positive while good defense is more negative.
Although DVOA is good for rating the average performance of a team/individual, it has a tendency of devaluing quantity. For example, if Jay Ajayi averages 4 yards per carry on 200 attempts, while DeAndre Washington averages the same on 20 attempts, it’s obvious that Ajayi is more valuable than Washington. However, DVOA fails to always account for this and will tend to give replacement players with small spurts of success the same rating as players who consistently perform at a high level. This is why DYAR is needed to put into perspective a high usage player’s performance. DYAR, or defense-adjusted yards above replacement, is effectively the total success points that the team/individual differs from a generic replacement team/individual. Thus, the higher the better.

This analysis will compare conventionally used statistics against DYAR and DVOA, as well as employ more qualitative analysis to understand whether these two advanced metrics are effective for comparing players and teams.

 

The Killer B’s

First, we need to take a look at whether or not the Steelers truly do have an elite trio. By diving into last season’s statistics, we can more or less extrapolate how good the Steelers’ trio has been over time, given that they have been playing at or near this level for most of their tenure together.

Tony T

Everybody knows Antonio Brown. The wide receiver has been ranked as one of the best at his position for the past four years, and he has only been improving. Given his perceived talent, the Steelers focus on giving Brown all the opportunities he needs to succeed. Brown fills up the stat sheet with 26.3% of Pittsburgh’s receptions. The team’s second most targeted wide receiver, JuJu Smith-Schuster, only registered 58 receptions for 15%. Given Brown’s usage, his eye-popping cumulative statistics (1,533 yards for most in the league) become less reliable, and we’ll have to take a look at advanced stats.

Taking a look at DVOA and DYAR, we can observe that Antonio Brown is well above over 84% of the league in both. Brown boasts the highest defense-adjusted yards above replacement, and a DVOA above one standard deviation away from the mean.

Figure 1 Red error bars represent one standard deviation away from the mean

Note however, the personnel that make up some of the top spots on this chart. JuJu Smith-Schuster solidifies his place with the highest DVOA and the Detroit Lions wide receiver Marvin Jones comes in with the second highest DYAR and third highest DVOA. Although these two receivers are definitely top notch, they are nowhere near the elite level of receivers such as Brown or Julio Jones who are able to consistently output good numbers. This makes it apparent that the DYAR and DVOA metrics are not very effective when grading wide receivers as they fail to isolate the effectiveness of the receiver and tend to muddle elite receivers with more replacement-level players.

Now we’ll consider two of the most telling stats for wide receivers – catch percentage and average air yards. Given that receivers who run shorter routes should definitely have a higher catch percentage, while those who run longer runs can have catch percentage of something closer to 50% to be effective, we can combine these two statistics allows us to discern how a receiver performs for his role on the team.

Figure 2 Average air yards per reception versus their catch % for receivers with 50+ receptions

The trend line represents how the average player performs in any given role. This metric appears to be much more effective, as the receivers we expect to see above the trend take their place. That being said, these metrics also fail to take into account usage and reliance. For example, Ted Ginn of the New Orleans Saints appears to be the biggest outlier, with 12 average air yards and a ridiculous 75.71% catch rate. That being said, he only has 53 receptions and generally is only thrown to when he has 3.5 yards of separation (the highest in the league). On the other hand, someone like DeAndre Hopkins gets thrown to with a measly 2.1 yards of separation, the second lowest in the league, and is still expected to make the catch. A receiver that is trusted to make the catch with fewer yards of separation is generally more reliable and more skilled overall.

Even so, these statistics do tell us that even with Brown’s extremely high usage he is able to consistently output success. Maintaining the highest catch percentage in the league for air yards over 12.5 is a huge feat for someone who is the team’s consistent number one option. The same can be said for his DVOA, which at 20.1% is only slightly above one standard deviation away form the mean (i.e. only slightly better than 68% of the league); however, achieving this value across so many receptions is what gives him such a high DYAR and makes him such a threat in the league.

In order to successfully represent every single one of these statistics (receptions, catch%, etc.), we decided to throw them all together and see what our output would be like in the following formula:Namely, wide receiver rating (WR) is determined by target air yards (A), catch percentage (C), separation (S), and targets (T). Based on our arbitrarily determined statistic, it comes as no surprise that Brown rolls in at second with 53.2. Interestingly enough, Hopkins tops the list with 62.17 and Julio Jones comes in third at 52.8. In any case, it is clear that Antonio Brown excels in every statistical category for wide receivers, and has earned his place as a top three wide receiver in the league.

 

Le’Veon Bell

Well known for his revolutionary patience when running, the Steelers young running back has been slowly showing himself to be one of the best running backs in the league. Rated by many to be the best running back in the league, Bell’s unique skillset has effectively changed the way many players and critics view rushing, with his patience behind the line allowing him to find clear paths to gain huge yards. That being said, is his style of play all hype or does it really produce results?

Bell is far and away the most used running back in the league, with 321 runs. For perspective, the second most number of runs is only 286. Compared to Antonio Brown’s usage, Bell’s is off the charts, and his total stats are therefore much less impressive. If we once again take a look at DYAR and DVOA, the statistics seem to do a much better job of ranking running backs. There appears to be a high correlation between average yards per attempt and DVOA, as Alvin Kamara and Dion Lewis rank extremely highly in both (6.1 and 5 respectively for average yards per attempt). The rating still fails to account for magnitude of attempts, as Bell more than doubled Kamara’s attempts, making his average gain of 4 yards much more impressive than it normally would be. Remember that attempts take energy, and being able to run the ball 321 times in a single season is a feat in and of itself.

With that in mind, Bell’s lower DVOA seems more reasonable due to his high usage. However, his extremely high usage actually makes his DYAR much less impressive. Given that DYAR represents total yardage above replacement, it should be the case that Le’Veon Bell’s DYAR should at least be top three with how often he has the ball.

Although DVOA and DYAR are much better statistics for measuring running back success rather than wide receiver success as there is no quarterback throwing them the ball when rushing, there is still an offensive line that is difficult to factor in. As such, these statistics are still somewhat subjective to the ability of a player’s teammates. Instead, lets take a look at success rate, one of the most important statistics in grading running back efficiency.

Figure 3 Top 30 players in yardage sorted from most to least

 

We can see here that Bell clearly beats out the majority of his peers in success rate. He sits tied with Marshawn Lynch and Mark Ingram, and is only beaten out by 6 other running backs in the league. On top of his high production, Bell is also capable of getting the results that his team needs. Note also that the success rate chart matches up quite well with the DYAR/DVOA analysis. Besides outlier Devonta Freeman, the same backs with elite success rates find themselves with extremely DYAR/DVOA values.

Beyond Bell’s rushing game, he is also the most targeted running back. With 85 receptions for 655 yards, Bell consistently provided Ben short flat or drag options to jump start the offense. That being said, Bell’s production on the receiving end was outclassed by a few of his peers. Alvin Kamara had fewer receptions but had over 150 more yards than Bell. Bell dominates the short game with both his running and receiving, but in terms of versatility he is not quite as elite as most people are led to believe.

Figure 4 Running backs with a minimum of 50 receptions

Bell has been often lauded for his high usage, with multiple 300+ carry seasons and a ridiculous number of total running yards. That being said, he fails to distinguish himself from other top tier running backs when it comes to general production. He averages only 4 yards per carry and consistently ranks below backs like Gurley and Kamara in every statistical category besides totals. His high usage gives his lower averages more leeway, however at a certain point it begs the question: is Bell’s usage a product of his performance or of the Steeler’s lack of other productive backs? All this being said, Bell is still a good player. He is simply not elite, and not quite as good as the rest of the world makes him out to be.

Big Ben

Off the bat it should be made clear that Roethlisberger is a Hall of Famer. With three championships to his name, he joins the era of elite quarterbacks alongside Drew Brees, Aaron Rodgers, Peyton Manning, and Tom Brady. That being said, Pittsburgh’s longtime quarterback is starting to get old and many are worried that his age is starting to catch up to him.

Taking a look at QB rating, out of all the starting quarterbacks who played in 13 or more games, Big Ben sits at a strong 7th. This puts him just on the cusp of the top-level quarterbacks currently in the league. To put this into perspective, below is a chart showing all quarterbacks with over 12 games, with a few of Ben’s peers highlighted.

Figure 5 As always, the inclusion of Flacco is questionable

In terms of quarterback production, adjusted net yards per attempt provides a strong metric when evaluating the ability of a quarterback to drive an offense. Comparing QB rating vs ANY/A (adjusted net yards per attempt) allows us to determine how effective a quarterback is in working a successful drive outside of gaining the most yards possible (similar to the success statistics for running backs). In this case, Big Ben does exactly what is expected of him in making the right play while also trying to maximize yards. However, being good is very different from being great. The QBR metric shows that Roethlisberger is a good quarterback, but he is outside the conversation for top-tier.

One caveat to these statistics is the difficulty in separating quarterback skill from receiver skill as we saw when taking a look at Antonio Brown. Using yardage metrics in general makes it difficult for quarterbacks such as Marcus Mariota, whose system relies more on the short passing game, to place as highly on the charts. In the case of Big Ben, Antonio Brown is very good at running deeper routes and this contributes greatly to Roethlisberger’s success. This also may be why players like Case Keenum appear to be elite quarterbacks, because they work with talented receivers such as Stefon Diggs.

DVOA and DYAR will hopefully be able to isolate the skill of individual quarterbacks. Taking a look at these metrics, we can see that Roethlisberger is considered at least top-tier, if not elite.

Interestingly enough, among the top performers in DVOA sits Jared Goff, one of the consensus top quarterbacks who had an abnormally low QBR. This begins to make more sense when we see that Goff’s number one passing options were Sammy Watkins and Robert Woods, two non-exceptional receivers. Furthermore, someone like Tyrod Taylor, who is considered by many to be an extremely mediocre quarterback, performs below average according to these advanced metrics, as opposed to QB rating where Taylor ranks 12th. It is safe to say that DYAR and DVOA might paint a better picture of quarterback skill than even the universally accepted quarter back rating. With this in mind, Roethlisberger’s 2017 performance would have to be considered top-tier and possibly elite.

Our analysis above allows us to conclude that the Steelers do in fact have a formidable offensive unit. Antonio Brown sits firmly as a definitive top three receiver while Roethlisberger defies age and continues to produce at an elite level. Bell’s statistics fail to matchup with his perceived top three running back skills, and land him just outside the conversation for elite backs. Even still, it is apparent that this offensive unit has all the tools to be ridiculously successful, even without accounting for one of the most clutch kickers in the game. Looking back at the statistics from past years, it is clear that the Steelers have had this high-level of personnel at least since 2013. So now, the question becomes, is the Steeler’s defense to blame or is there something that individual statistics can’t tell us?

 

Team Metrics

The NFL is a very high variance league. With playoff games being one and done deals, predicting playoff success is extremely difficult no matter what kinds of statistics are at your disposal. With that in mind, take the following analysis with a grain of salt. Even though the Steelers have seemingly underperformed in the playoffs the past few years, it is only because the expectations for this team are sky high. It is entirely possible that the team has been on a rather unlucky stretch for the past few years.

In terms of offensive production in the regular season, it comes as no surprise that Pittsburgh ranks 3rd in offensive DVOA. On top of that, their schedule puts them right around league average, with the 18th hardest schedule in the league this year. Past statistics show similar results, with the Steelers in the top three for offensive DVOA in both 2015 and 2014 as well. Their worst year of 2013 had them at 12th due to one of the worst rushing DVOA’s in the league. This coincides with Le’Veon Bell’s initial breakout year. With that in mind, we’ll be focusing on the 2014-17 seasons.

By most metrics, Pittsburgh has the statistical prowess to be a consistent championship caliber team like the Patriots. Diving straight into DVOA, the statistics show us that in general, championship caliber teams will have an extremely strong offense (+10%) and an above average defense. Below is a graph showing teams that have made the conference championships since 2014.

Figure 6 Conference championship teams since 2014. Super bowl winners are in red

The triangles represent the Steelers, with their statistics in the past four years matching up very well with those of the championship teams. Moreover, when we isolate the DVOA statistics for the Patriots, who are represented by squares, a clear trend appears, albeit a small sample size. The chart highlights the strong similarities between the offensive and defensive DVOA of the Steelers and the Patriots. Both even have a single outlying year each. With the very apparent success of the Patriots, it seems as if the Steelers are missing something that cannot be represented by stats.

Taking a look at more nuanced data, a startling trend appears for the Steelers’ conversion rate. Although the Steelers have consistently had decent 3rd down conversion rates, even finishing 2nd in the league in 2017 for 3rd down conversion percentage, the team has a very surprising inability to convert 4th down attempts and red zone scoring opportunities. Moreover, this lack of conversion success appears to be one of the major differentiating factors between the Steelers and the Patriots.

Figure 7 League rank for 4D% vs Red zone % between 2014-2017

Above is a chart displaying league rank for 4th down conversion percentage and red zone conversion percentage for the given team. That means the closer the team is to zero, the better. New England consistently shows a high ability to convert clutch plays while Pittsburgh seems to be all over the place, often falling near the bottom of the league in one or the other category. The one instance when the Steelers had clutch conversion rankings similar to that of the Patriots, they were in the AFC championship game.

Now we know that the Steelers’ have had historic issues converting when it mattered. How exactly do these stats play into post-season performance?

Taking a look at this past postseason in particular, the results are strikingly apparent. In the game against Jacksonville, only one Pittsburgh Steeler score came from within 15 yards of the end zone (not including the final drive which ended in the last second of the game). In comparison, the Jaguars had five touchdowns as a result of red zone plays, and the rest of their points were scored from a field goal and a fumble return. Going back to 2016, against the Kansas City Chiefs, the Steelers failed to score a single touchdown, with every single one of Chris Boswell’s field goals coming from within the red zone.

Not only are the Steelers unable to convert clutch situations, they also tend to not attempt plays on 4th down anyway. This passive style play is another contributing factor the Steelers’ lack of success. Playing it safe on 4th down is usually detrimental, and the Steelers do so consistently more than other teams in the NFL. The Steelers were ranked 4th, 8th, and 3rd in least number of fourth down attempts outside of 2016, when they were ranked 19th and made it to the conference championships.

In addition to poor conversion rates, it is also quite possible that the Steelers have simply been over-performing during the regular season. Through the past 4 seasons, the Steelers have had a consistently lower expected Win-Loss ratio than their actual record (where expected win-loss is calculated based on point differential). In a historically weak AFC North division, it comes as no surprise that the Steelers would perform well in the regular season. Every season, the Steelers get lucky and blow up the league with a great regular season record. But come postseason, the boys in black and gold revert back to the mean, and lose the games they should be losing.

Obviously there are various other factors outside of clutch conversions and expected win-loss which differentiate the Steelers from the Pats. The coaching staff, the style of play, and even the team cultures are wildly different. That being said, there are clear aspects of Pittsburgh’s game that clearly need work. Even with a lineup as intimidating as the Killer B’s, the Pittsburgh Steelers are not the team people think they are. Defense-adjusted stats show the Steelers as one of the best offenses in the league, but only when unchallenged. In the red zone and on 4th downs, Pittsburgh fails to perform, and based on their usual point differential, it makes sense. This just goes to show; a handful of players don’t make a team. It takes more than just three guys to get a team to the championships. With Le’Veon Bell likely leaving the team in the near future, Pittsburgh will be forced to reevaluate their position in the league and perhaps finally figure out what’s been missing from their championship caliber team.

Leave a Reply