Tag: EPA

High Impact: Ezekiel Elliott still foundational to Cowboys’ offense

Feed Zeke is gaining more evidence as the Cowboys have shifted to a pass-dominant philosophy in 2020.

The Dallas Cowboys did it again. There are two early themes to the Cowboys’ 2020 season, injuries and overcoming double-digit deficits. On Sunday, they dealt with both in a major way. Dallas fell behind the New York Giants 17-3 in the blink of an eye, came back, lost their starting QB for the year to fall behind again, and came back again.

The last-second victory brought the Cowboys record to 2-3. Despite the heart attacks and heart ache that seems to be the central theme of this season for the Cowboys, momentum-swinging plays continue to define things. In a bit of contradiction to this idea, homage has to be paid to the effort of Ezekiel Elliott for providing the non-chunk plays which set the pace.

Ending the game with two scores and 125 yards from scrimmage, once again when his team needed juice it was Elliott who steered the ship for the Cowboys.

Trailing 17-3 early in the second quarter, Dallas started with the ball at their own 25. To that point, Elliott — who no longer is the central focus of the club’s early-game plans — had three carries and two targets out of 16 non-kick snaps. His carries went for seven yards, his one catch was for 14 yards and the other target resulted in a pick-six when Prescott’s throw sailed over his head.

But just like the Week 2 home contest against Atlanta — incidentally the Cowboys only other win in 2020 — turning to Elliott is what got Dallas calmed down when their offensive world was spinning out of control.  In Week 2, Elliott’s fumbles were part of the malaise, but the team started out with two runs by Tony Pollard and was immediately out of sorts.

Trailing 20-0 at the end of the first quarter, the club turned to Elliott to settle down. On their first scoring drive of the game, the former first-round pick saw the rock on 7 of 11 snaps as the club methodically marched down the field.

  • Pass to Cooper for 11 yards
  • Elliott up the middle for 8 yards, plus defensive holding
  • Elliott right guard for 7 yards
  • Pass to Lamb for 9 yards (end of quarter)
  • Lamb right end for 9 yards
  • Shotgun run by Elliott for 10 yards
  • Pass to Elliott for -3 yards
  • Pass to Bell for 10 yards
  • Elliott for 7 yards (on 3rd and 3) injuring tackling player
  • Elliott for 1 yard
  • Elliott for 1 yard, Touchdown, injuring tackling player
(AP Photo/Ron Jenkins)

The Cowboys after going punt, fumble, fumble, turnover on downs, fumble then went on to score on their next four possessions and seven of their final eight drives.

On Sunday, Elliott saw the ball on 9 of the 14 plays when down 17-3.

  • Elliott off right tackle for 9 yards
  • Elliott up the middle for 5 yards
  • Pollard for 5 yards
  • Pass to Cedrick Wilson for 4 yards
  • Offside on Giants
  • Elliott right tackle for 3 yards
  • Elliott left guard for 8 yards
  • Pass to Wilson for 10 yards
  • Elliott right tackle 3 yards
  • Pass to Gallup 8 yards
  • Elliott right guard for 8 yards
  • Elliott left tackle for 1 yard
  • Pass to Wilson for 5 yards
  • Elliott for no gain
  • Elliott for 1-yard touchdown

15 snaps, one penalty and 9 touches for the workhorse back. That jumpstarted Dallas who finally got to Giants’ QB Daniel Jones the next drive and scored their first defensive touchdown of the season. They followed that up with another offensive score for three straight touchdowns and took control of the game.

(AP Photo/Ron Jenkins)

There’s no question Dallas is 100% right by becoming a pass-heavy team. There’s no question the Cowboys are more difficult to defend when they have diversity on early downs and skew towards the pass. However, there’s also no question their offensive identity for the last four years is in Elliott being the centerpiece and going away from that seems to have the offense in disarray over the first five games of the season.

In the games where the Cowboys have been able to start fast, or at least reasonably well, Elliott has been heavily involved from the outset. That includes the season opener against the Los Angeles Rams, where the teams were tied at 7 early in the second quarter, and Week 4 against the Cleveland Browns, where Dallas led 14-7 going into the second quarter.

In the small sample size of NFL seasons, five games is even more minuscule, but it certainly looks like the Cowboys are better off starting games when Elliott is involved, and he’s been the reason they’ve found their footing in the games they weren’t.

The game against the Seahawks in Week 3, Dallas started off with six straight passes, so even with Elliott getting adequate work on the drive with three touches, things still felt off from the jump.

Advanced metrics allow us to know so much more about what the proper decisions are in football. Expected Points Added, or EPA, is one of the leading indicators as it says on a play-by-play basis how much a decision contributes to the possibility of a score. It’s the foundation of the running-backs-don’t-matter movement, along with the obvious evidence that running backs more than wideouts and quarterbacks can be found at later points in the draft.

But we’ve long since argued Elliott is outside of those parameters for multiple reasons.

He is continuously among the best of his position, and clearly, the Cowboys offense centers around him.

This year, EPA seems to match the assessment. Passing will always dwarf rushing when it comes to EPA, but in looking at the Dallas offense compared to the rest of the league, they have comparatively been a better rushing team than passing.

Despite still being a net negative (-0.011) Dallas ranks eighth in the league in rushing EPA; they rank 19th in passing EPA (0.115) despite Prescott being on a record-setting passing-yard pace prior to his injury.

Despite Elliott’s three fumbles, the Cowboys run game has been highly productive.

Data science has claimed to prove that passing success is independent to having defenses need to respect an opposing running game. But when Dallas still relatively skewed run in 2019, the passing game EPA was through the roof. Now that the Cowboys skew pass heavily, the EPA has dropped and the run EPA is among the league best.

Again, small sample size may have a lot to do with it, but it’s certainly more than anecdotal the Cowboys in 2020 have fared better when Elliott is a focal point.

Feed Zeke.

Watch: Catch This Fade, S5E12

Subscribe here on Patreon to get full eps, audio and video tiers available

[vertical-gallery id=656015][lawrence-newsletter]

Good, Bad, Ugly: How Brett Maher’s ineptitude submarined Cowboys yet again

A breakdown of the good, bad and mostly the ugly from the Dallas Cowboys’ crushing 31-24 loss at the hands of the Chicago Bears on Thursday.

The Dallas Cowboys’ crushing 31-24 loss to the Chicago Bears on Thursday night was a signature game. The Cowboys have gone through various ups and downs during the 2019 season, but they are now riding a three-game losing streak at the hands of the New England Patriots, Buffalo Bills and now the Bears. Technically all is not lost as the Cowboys still hold first place in the woeful NFC East, despite their middling record of 6-7.

It has been an unfortunate turn of events for the Cowboys. The expectations of head coach Jason Garrett leading a star-studded roster into a Super Bowl run now seems like a pipe dream as the team again looked outmatched in every aspect. It was a bad scene and at times it looked liked many players were on the field reluctantly. Here is a breakdown of the good, the bad and mostly the ugly from Thursday’s primetime defeat.

The Good: The Opening Drive

David Banks-USA TODAY Sports

The Cowboys looked like a dominant football team to start the game. Their opening drive was impressive as they gained 75 yards on 17 plays. It elapsed 8:57 of game time and the end result was a rushing touchdown from Ezekiel Elliott for an early 7-0 lead. It was the longest drive of the 2019 season and by the time the Bears got the ball on their first possession there was only 6:03 left in the first quarter.

The Cowboys effortlessly moved the ball at will, and if the opening drive was to be an indication, Cowboys Nation was sure in for a treat on Thursday night. As it would eventually turn out, that wasn’t the case. In truth, there is not many positives to gleam from the loss. Playoff hopes and scenarios aside, it seems like the product on the football field has not matched the hype.

The Bad: Brett Maher

Cowboys’ owner and de facto general manager Jerry Jones has already voiced his opinions about Garrett and the state of his football team. At this point, it is safe to say the entire coaching staff will be allowed to finish up the season uninterrupted with the team’s current playoff positioning. With that fans can likely, sadly expect Garrett’s continued dependency on kicker Brett Maher in crucial situations.

In fairness to him, he has made some long kicks in his two-year tenure. But a couple 60-plus-yard field goals do not make up for the plethora of misses that have plagued the team throughout the season.

For the second-straight week, Maher’s misses had a direct impact on Dallas’ late-game decisions, keeping them more than one-possession down late in the contest when they were finally able to put points on the board. The morale erosion when the offense makes headway only to come away not just empty handed, but gifting the opponent prime field position resonates throughout the sideline for a team barely hanging on to belief they can play a complete game.

To make matters worse in this contest, after Dallas was able to shrink the lead to 24-14 with a touchdown, Maher’s kickoff dribbled out of bounds, starting the Bears off at their 40-yard line.

The Bears went on to score in four more plays, their final points of the game and the final margin of victory; set up by the short field.

The conservative nature of Garrett and the inaccuracy of Maher has been the perfect disaster for a Cowboys team struggling to win in close games. But with three games left in the season the misses will continue to be a problem. Much like their current coaching situation, ownership seems adamant to stick with Maher for the remainder of the season in spite of the mediocrity. The Cowboys’ season in a nutshell.

The Ugly: The Cowboys’ Defense

AP Photo/Morry Gash

It was another poor outing for a porous Cowboys defense. They allowed the Bears, led by quarterback Mitchell Trubisky to score 31 points. In addition, the defense had trouble tackling all night, which led to some bonus yards after the catch opportunities for the Bears’ offense.

To put things into perspective, the Cowboys’ defense allowed Trubisky, who is an average QB at best, to pick them apart with surgical precision. To gauge Trubisky’s season from an analytical standpoint, ESPN’s Total QBR is a good measuring stick. According to their metric, Trubisky currently ranks No. 27 in the league with a a QBR of 43.0 for the season. A Pro-Bowl caliber player is rated with a QBR of 75. Against the Cowboys Trubisky produced a QBR of 80.6.

Another was to measure Trubisky and the rest of the Bears’ offensive efficiency is through the use of Expected Points Added (EPA). The nflscrapR package provides this data, with EPA measuring the value of a play given down, distance to first downs and field position.

This chart illustrates the Bears’ offensive efficiency, or lack thereof, over the course of the season. For the most part, the Bears have struggled to produce on the offensive end of the spectrum. Their EPA/play was usually well below zero, which means that the majority of their plays were unsuccessful.

Against the Cowboys in Week 14, they saw one of their biggest offensive breakouts of the season. The Cowboys’ defense surrendered 0.36 EPA per pass play, which is the Bears’ greatest mark of 2019.

Overall, this was a daunting loss for the Cowboys who were the road favorites against a Bears team led by Trubisky. There was hope the Cowboys would bounce back in Week 14 after their stunning defeat by the Bills on Thanksgiving. The team had a chance to prove they were a resilient bunch, able to withstand a string of losses. Now, after a three game losing streak and a record of 6-7, this team is treading the same old ground to end the 2019 season.

[vertical-gallery id=635660][vertical-gallery id=635755][lawrence-newsletter]

Even Jason Garrett can use these analytics showing who the Bills really are

A look into how the Bills have accomplished their 8-3 record, whether it’s fool’s gold or legitimate, and what Josh Allen is doing so well.

Despite it being the week of Thanksgiving, Sunday’s result didn’t leave Cowboys fans with a feeling of gratitude. The field became a quagmire in the midst of a nasty rainstorm, and it made life difficult for both quarterbacks and offenses. To this point in the season, the Dallas offense has been more reliant on Dak Prescott’s ability to make plays, and thus the weather better served the New England defense as it mitigated Prescott’s potential impact.

Games like Sunday’s contest demonstrate the importance of having every possible edge a team can find. Uncontrollable elements like the weather can level a matchup in unexpected ways, and thus its important to maximize the value of your own decision-making. In light of this loss, and perhaps with the rise of the Baltimore Ravens, Jason Garrett has come under scrutiny for the lack of usage of analytics in the team’s decision-making. Baltimore is the latest team to grab headlines for this usage, but New England has been at the forefront for the majority of their dynastic run and now Dallas will face another analytically-inclined franchise in the Buffalo Bills.

Buffalo comes to town holding an 8-3 record, and the top Wild-Card spot in the AFC. Due to historic-level production, the New England and San Francisco defenses have garnered the attention of most headlines. But the Buffalo defense has played a significant role in their record to this point.

The truth, however, is that they’ve likely benefited from some weaker offenses.

Here we see the EPA per play that a defense has allowed (Y-axis) vs the EPA per play the offenses generated against other teams (X-axis).

The idea is to compare a team’s defensive production against their quality of competition.

And as you can see, Buffalo’s opponents have been quite abysmal overall. In fairness, Buffalo has performed well against those opponents. But its safe to say that their defense has been largely untested. The Dallas offense should theoretically give a better indication of what Buffalo’s defense is capable of achieving.

Because their opponents have been so appalling, it’s hard to glean what’s real from Buffalo’s defensive statistics. However, a player worth noting is cornerback Tre’Davious White and his lock-down potential.

According to Pro-Football-Reference, White has allowed a 47.9 passer rating on passes his way, ranked second among all cornerbacks. He’s a player who’s shown the capability of shadowing a team’s top wide receiver, so it should be interesting to see how Buffalo chooses to deploy him.

On the opposite side of the ball, it’s hard not to talk about the Buffalo Bills without talking about their young quarterback, Josh Allen. It seems fitting Allen comes to town in a week where analytics was featured as a prominent discussion point. Entering last year’s draft, Allen was the prototypical case in skills vs stats. From a physical standpoint, it’s hard not to be wowed by his talent, as he blends an unprecedented amount of arm strength and mobility. Despite this though, things never quite came together at Wyoming, and many wondered whether Allen’s physical tools would translate to success on the NFL field.

Allen put these physical tools on display immediately as a rookie, but we also saw his rawness play a role. That said, Allen’s development as a quarterback is rather impressive when you observe his Next-Gen passing charts from 2018 and 2019.

In his rookie campaign, it was clear how much Allen believed in his arm, as he routinely attacked downfield, and found great success attacking the seams. He did, however, struggle to attack the perimeter of the field. But in 2019, we see Allen utilizing the entire field, without really sacrificing much of his downfield shots. It’s almost downright laughable that Allen has the ability to uncork a pass to any point on this chart. As a defense, it means having to defend every level on a given play, and can prove costly if there’s a lapse in focus.

Arm strength of Allen’s caliber is certainly rare, but it’s the combination of strength and mobility that truly sets Allen apart. Common sense would dictate players should be keenly aware of their strengths, and stick to those as much as possible. This is certainly the case with Allen, as no quarterback has relied more heavily on his legs this year.

The chart above shows that about 12% of Allen’s successful plays have come due to his rushing ability, while about 9% of his dropbacks resulted in a scramble overall. While Allen has shown development as a passer, he still likes to use his legs, and can certainly do so if unaccounted for by the defense.

Buffalo’s schedule to this point may have inflated their overall record, but they’re still dangerous enough to cause trouble for a team, especially on a short week. Dallas will need to regroup quickly if they hope to avoid an unpleasant Thanksgiving. But with a holiday game, the eyes of America will be tuned in, and that presents a unique opportunity for Dak Prescott to vault forward in the MVP discussion.

[vertical-gallery id=635321][vertical-gallery id=635041][vertical-gallery id=634743][lawrence-newsletter]

Brady vs Prescott: A one-sided battle is on tap in Cowboys-Pats

A look at the premium position in Sundays matchup between the Dallas Cowboys and New England Patriots.

People love sports for different reasons. Sometimes, it’s the passion or the emotion associated with it. Families bond around Sunday afternoon victories (or wallow in the losses), and memories forge loyalty unlike anything else we experience. For me, I always liked football for its passion, but it was the mental aspect that made me fall in love.

The NFL is a 32-team chess match. This metaphor doesn’t just apply to the in-game strategy on Sundays, but the overarching philosophy a team chooses to deploy in roster construction and team-building as well. It means that the league is non-static, as each organization searches for the key that unlocks the next wave of success. That’s what makes this current season so interesting. Players like Patrick Mahomes, Lamar Jackson, Deshaun Watson, Dak Prescott and Russell Wilson are completely changing the landscape of what quarterbacks are capable of achieving. From a historical perspective, the current level of quarterback play is unprecedented.

At this point in time, there are several advanced metrics we can use to quantify the production of a team’s passing attack. I took seven of these metrics and made the effort to condense their information into two metrics. The rationale here is that we can combine information from the different stats to get an overall glimpse at player production (and two metrics fit into a graph more easily than seven). If you’re versed in statistics, this was done using principal component analysis, and the seven metrics used were:

  • EPA per Dropback
  • Success Rate (percentage of positive EPA plays)
  • Average Depth of Target (ADOT)
  • Sack Percentage
  • Interception Rate
  • Adjusted Net Yards per Attempt (from Pro-Football Reference)
  • Completion Percentage Over Expectation (CPOE)

In doing this, a significant portion of the signal was condensed into the first two components. Here’s a graph of single-season QB performance dating back to 2009, filtered to show passers with at least 100 pass attempts on the season.

A little bit of information on how to interpret these composite scores.

The first component (X-Axis) is an overall indicator for passing performance. In essence, it rewards a player for the good stuff (EPA, CPOE) and docks them for the bad (sacks, interceptions).

The second component (Y-Axis) isn’t quite that simple.

Here, this isn’t so much an indicator of quality as it is style. In a way, this composite score is capturing the aggressiveness of a passer, as it scores higher for players with a higher ADOT and interception rate, but penalizes them for a high CPOE.

I’ve highlighted the quarterbacks for this week’s matchup.

In the upper-right corner we see Dak Prescott’s 2019 season. In terms of his first composite score, Prescott ranks third overall  just behind Aaron Rodgers’ 2011 MVP campaign and Drew Brees’s 2018 MVP runner-up year (neither highlighted).

There are several other 2019 quarterbacks in the upper echelon of the first composite which would seem to suggest that teams are now better at leveraging a player’s skills for passing efficiency than they were in years past.

But encouragingly, Prescott is performing at this level while maintaining an aggressive approach, a positive indication for the explosive capabilities of the Dallas offense.

On the flip side, we find Tom Brady’s 2019 campaign just slightly above the overall average. As the face of the NFL and backbone of an epic dynasty, we’ve grown accustomed to Brady orchestrating the New England offense to fantastic efficiency, but the 2019 campaign has not reached those levels of expectation.

However, if we look at how this first composite score correlates from one year to the next, Brady is actually performing right about where we would expect, based on last year’s results.

Conversely, there’s been a common theme from many that they didn’t expect to see this caliber of progression from Prescott, and his first composite score would echo this sentiment as he’s performing well above what we’d predict given historical trends.

But if we revisit that first graph, we also see Brady scores relatively low on the second composite score.

If we take a look at his air yards distribution this season, it’s not hard to see why.

Brady is attempting passes of short length at a rate higher than the rest of the NFL, and as a result he’s attacking the deeper part of the field less frequently.

To be specific, 56% of Brady’s pass attempts have come within 5 yards of the line-of-scrimmage. The retirement of Rob Gronkowski and a revolving door of receivers have certainly played a factor here. Another way to see this is to compare the completion percentage density charts of Prescott and Brady.

We can see here the vast majority of Brady’s targets are around the line-of-scrimmage, and naturally he completes them at a high percentage. However, we also see that he hasn’t found great success targeting the perimeter, especially beyond 10 yards. When Brady has found success down the field, it’s most commonly been down the seam.

Prescott’s 2019 season seems all the more impressive in this chart, as there doesn’t appear to be an area of the field he can’t own. However, the chess game continues on Sunday, as Prescott faces his toughest challenge in grand-master Bill Belichick.

Here’s hoping the evolution of Prescott and the NFL continues.

[vertical-gallery id=634743][vertical-gallery id=634702][vertical-gallery id=633628][lawrence-newsletter]