clock menu more-arrow no yes

Filed under:

Windy City Gridiron Fact Check: Trubisky and the Deep Ball

New, comments

Every so often, it’s nice to have a fact check on some of the opinions and speculation running around about the Chicago Bears—especially when quarterbacks are involved.

Chicago Bears v San Francisco 49ers Photo by Thearon W. Henderson/Getty Images

A number of people have opinions on Trubisky’s deep ball. For example, I have expressed concerns over his hip motion and alignment when throwing the ball, especially from the pocket. However, the problem with these sorts of opinions is that they frequently rely on anecdotes and one-off observations. A single impressive connection or a brutal miss can weigh heavily in our minds, and that can influence what we see (or think we see) the next time through.

1520 passes traveled over 20 yards through the air in 2018 over the course of the 256 regular season games and the playoffs. That means that each game really only witnessed five or six long passes. A “typical” quarterback might throw 40 to 45 such passes over the course of a season, or a handful per game. Two misses or two completions in a row could linger in a fan’s mind, ossifying an opinion, when they were really just anomalies.

Therefore, this article aims to look at the numbers available for Mitchell Trubisky’s deep passing ability and to put those numbers in context. Similar work has been done by Johnathan Wood, a member of the WCG Community who also writes for Da Bears Blog. However, for this piece I want to use as much independent analysis as possible.

Before diving any deeper, I should point out that I am heavily indebted to Pro Football Reference, Football Outsiders, John Kinsley, and NFL NextGen stats.

Let’s begin.

Background: The Deep Ball Project

For the 2017 season, John Kinsley wrote a piece for Football Outsiders attempting to quantify the deep ball, and Kinsey writes for a separate website where he maintains the Deep Ball Project. One of the great things about Kinsey’s work is that he separates “completion percentage” from “accuracy.”

For example, his numbers exclude drops, plays where the receiver fails to maintain control of the ball through to the ground (and plays where the ball is knocked from the receiver’s hand before he can complete the catch), and other circumstances where the pass was accurate even if the play was incomplete. He also tracks things like inaccurate completions--plays where the ball was a completion but it was because the receiver adjusted to an inaccurate pass.

During his rookie campaign, Trubisky’s 50% accuracy percentage on passes of over 16 air yards was enough to earn him 13th place out of 35 qualifying quarterbacks. The average that season was 48.1%, though, so it’s not like his numbers strayed significantly from the mean. Likewise, out of 52 qualifying attempts, the Bears’ quarterback had 4 inaccurate completions--plays where his receiver bailed him out despite the ball getting away from him. This might lend some fuel to the fire that he might have benefited from his receivers’ ball skills, except this was back in 2017 when receivers with ball skills were hard to find in Chicago. Additionally, he also had exactly 4 “accurate incompletions,” plays where his pass went to the right place but other factors interfered. Overally, Trubisky’s efficiency score (using a metric that Kinsley explains much better than I could) was 0.96, compared to a league average 0.99.

In short, with sample sizes this small, Trubisky was a perfectly average quarterback his rookie year when it came to down-the-field accuracy. There was just enough inconsistency there to give his detractors and his defenders equal ammunition. He was 2nd in the league on passes going 25-29 air yards and fifth on passes going 35-39 air yards. He didn’t attempt passes at some ranges, and so Kinsley’s system penalizes him there, but his real weakness seemed to come on passes at 20-24 air yards (where he only completed 37.5% of his passes and came in 32nd).

Of particular note to me was the fact that a rookie Trubisky was 10th in the league when it came to throwing under pressure (accurate 38.5% of the time) but 26th in the league with a clean pocket (61.5%). So, prior to his first year with Matt Nagy, Trubisky was proving everyone right. He was able to throw a deep ball well when he was on the move, but in a clean pocket he’d make some really bad misses.

It’s also worth noting that Trubisky aired the ball out a lot in 2017, relative to his total number of passes. He had 52 qualifying passes out of 330 attempts, meaning he went deep on about 16% of his attempts. By comparison, two of the leaders in deep ball accuracy were Dak Prescott with 14% (68/490) and Aaron Rodgers with 12% (29/238). Thus, in each of his 12 games, he gave fans an average of four chances to form an impression, and those impressions were erratic enough to support any narrative.

Belief #1: Trubisky Can’t Throw Left

For the 2018 season, Kinsley was thoughtful enough to break down deep passes by direction. Of the 24 passes thrown by Mitchell Trubisky that were over 21 yards in distance, Trubisky was accurate on more than 54% of them. That was good for 7th in the league when throwing left, and only 5 quarterbacks went left more often than he did.

NextGen stats disagrees a little, in that they have Trubisky with a passer rating of 65.5 on throws to the left over 20 yards (the NFL average was 76.6). NextGen still calls this performance within average, and that’s probably a fair assessment once drops and tips and the rest are factored in. However, even penalizing him for throws that Kinsley excludes, he’s not terrible. When compared to the data from 2017, his deep left ball was actually good for a 118.8 rating--with an admittedly small sample size. He did struggle on the intermediate range throws to the left in 2017, but then did better than average in 2018.

If #10 struggled, it was with the small number of throws that were to the deep middle. He only attempted 8 such passes, and that might have been seven too many. He had an accuracy rating of 37.5% and a passer rating of 25 when throwing down the middle. However, there is little to no empirical data that shows a weakness to throwing to the left, and there is even some information to suggest that deep left balls are a possible area of strength.

Verdict: Implausible. Does Trubisky struggle with throws to the deep left sometimes? Yes, but no more so than other quarterbacks.

Commentary: I was legitimately surprised by this, but as someone who watched Jay Cutler play for years and as a tennis fan who remembers Andy Roddick’s service motion, I should not be. While I personally dislike Trubisky’s throwing motion when he throws to his left, an awkward motion does not automatically mean failure.

Belief #2: Trubisky Lacked Touch on Deep Balls

There are some players who excel at throwing into tight coverage, and those players tend to become truly dominant quarterbacks. One complaint about Trubisky’s deep passes is that they lack fine touch. The numbers do seem to back this up. In 2017 he had a 62.5% accuracy rating on open windows and a 39.3% accuracy rate on tight windows, but the discrepancy there is not shocking and it’s something many quarterbacks faced. However, the difference became more pronounced in 2018.

A lot of Trubisky’s best deep plays from last season happened when he had open windows--he was 17/24 (71% accuracy). This is likely the Matt Nagy effect, generating space around receivers to give Trubisky room to work. When that room isn’t there, Trubisky plummets to a 30% accuracy rating (12/40, or 23rd in the NFL). This discrepancy goes a long way to explaining the Air Yards Differential found by NextGen Stats. Trubisky ranked 30th of 39 qualifying quarterbacks for the difference between his completed air yards and his intended air yards.

Verdict: Plausible. Yes, Trubisky is completing downfield. However, when he does so, it seems to be because his receivers have space. When forced to fit the ball into a tighter window, he struggled in 2018 more than other quarterbacks.

Commentary: The large number of missed throws into tight windows (almost 10% of Trubisky’s passes) suggests that Trubisky spent 2018 playing aggressively. The fact that he tied for the 10th highest interception rate in the NFL (and 8th highest adjusted interception rate) suggests that less aggression might be better for the team, unless Trubisky can work on his downfield accuracy.

Belief #3: Trubisky Needs a Clean Pocket to Deliver Deep

This is one of the stranger takes I have read, because Trubisky is a gifted scrambler. However, I have read multiple accounts wherein people suggest that he only connects deep when he has a clean pocket.

NextGen isn’t much help here, but conventional DVOA at least serves as a basic crosscheck. Trubisky’s DVOA without pressure is 37.9%, only good for 26th in the league. Meanwhile, his DVOA when facing pressure is -14.1%, a pretty dramatic drop. However, that drop is usually even worse. In fact, under pressure, Trubisky’s DVOA was 3rd-best in the league in 2018. More than that, the not-quite 52% fall-off in DVOA under pressure was the lowest penalty suffered by any quarterback in 2018. As much as possible, Trubisky “thrives” facing pressure, at least compared to other quarterbacks.

Kinsley’s tracking backs this up. Trubisky had a clean pocket on 45 throws deep throws, and his accuracy was 42.2% (28th in the league when in a clean pocket). Meanwhile, he was under pressure on 19 deep throws, and his accuracy there was almost 53% (3rd in the league).

Only two quarterbacks had a higher number of deep passes from a clean pocket than Trubisky (Rodgers and Roethlisberger). Rodgers had qualifying deep passes on 70 of his 597 attempts (12%), and he had a clean pocket on around 9% of them. Meanwhile, Roethlisberger had qualifying deep passes on 62 of his 675 attempts (just under 9%), and he had a clean pocket for those deep throws on around 7% of his passes. Trubisky, meanwhile, had qualifying deep passes on 64 of his 434 passes (15%), and he had a clean pocket on 10% of them.

Now, it’s hard to draw conclusions from one year’s data, and I don’t know how many passes Kinsley disqualified, but there are two obvious possible inferences from this information. The first is that Trubisky liked to go deep, especially when the line was giving him the chance to do so. The second is that the line gave him an amazingly clean pocket last year to work with.

Verdict: Implausible. Trubisky operated better under pressure than he did in a clean pocket on the deep ball. There is, however, some evidence to suggest that he has a tendency to go deep when given the chance. There is also evidence that he needs to improve his accuracy from the pocket.

Commentary: It’s possible this offensive line is better than we think it is.

Bonus Belief #4: Trubisky Needed Yards After Catch to Convert

This is not purely a belief regarding deep balls, but it is a connected belief among fans and commentators. Simply put, there are those who believe that Nagy’s offense only worked by having Trubisky dump the ball off to an open receiver and then having that player make the conversion. This sort of “check down” play is tracked by Scott Kacsmar’s advanced statistic “ALEX,” described at Football Outsiders as follows:

For those new to this metric, it stands for Air Less EXpected, or ALEX for short. ALEX measures the average difference between how far a quarterback threw a pass (air yards) and how many yards he needed for a first down. If a quarterback throws a 2-yard pass on third-and-8, then that would be -6 ALEX. The best application of ALEX is to look at third down, when it’s really crucial to get 100 percent of the needed yards to extend the drive.

If Trubisky were to struggle to convert third downs, or if he consistently threw short of the chains while still converting, then this would be an indication that the belief that he relied on YAC might have some merit. Interestingly, in 2017, there is an indication of this trend. Then that trend changes. Kacsmar himself summarizes the change perfectly:

Mitchell Trubisky had the lowest ALEX (-2.4) as a rookie and he was 28th through Week 5 (-0.6). But under Matt Nagy, Trubisky changed his style to finish seventh (+2.2) in both ALEX and conversion rate. He also went from having the most need yards (8.8 yards) on third down under John Fox to the fourth-lowest average (6.9 yards) this season, so that always helps.

So, under Fox, he was throwing short of the chains, but he also averaged needing almost 9 yards when passing on third down in order to convert. That’s...brutal. It’s also enough to form a really negative impression in the minds of fans when he’s throwing almost one and half Tarik Cohens short of the line of gain. Interestingly enough, the ball was going more than 9 yards through the air when he needed to convert on third down in 2018.

As the trends noted above suggests, many of those conversions did likely come to open receivers instead of being made into tight windows. However, the throws themselves still had the distance, and were still being completed. In fact, the 2018 version of Trubisky ranked 13th in the NFL in converting third downs when a long pass (here defined as only 7+ yards) was required.

Verdict: Implausible. Trubisky was able to deliver the ball through the air adequately enough to convert at a high rate. He did not rely on yards after catch, at least on third down, even if the offensive system utilized by Nagy uses yards after catch as part of its overall system.

Commentary: There is a documented phenomenon known as the primacy effect. To put it simply, we tend to form our impression of things based on our early encounters, and it becomes increasingly difficult over time to change our impression. Those who believe that Mitch cannot connect on the deep ball, or cannot complete without his receivers doing most of the work, might very well be seeing what they expect to see after having that impression formed during 2017--when it was demonstrably true. However, it is more likely that someone’s “eye test” is in error if they believe they are still seeing that effect play out when the objective metrics suggest otherwise.