Examining Chicago’s Personnel Usage/Tendencies on Offense in 2020

Like I’ve done the last few seasons, I want to explore how the Bears deployed their skill position players on offense in 2020 to see if there are any trends or tells for which opposing defensive coordinators can look. These are tendencies Chicago’s coaches should be aware of and look to rectify in the future.

The table below shows changes in run percentage when skill position guys who played between 35-65% of the snaps were in the game vs. on the sideline.

On the high end, that excludes players who played more than 75% of snaps, because their “off-field” splits would be too small to consider. That was only Allen Robinson in 2020.
On the low end, that excludes players who played less than 25% of snaps, because they are often mainly in the game in specific situations, where a run or pass may be expected (i.e. the 4th WR in a 4 WR set for 3rd and long, or the 2nd TE in a short-yardage set). This excluded Demetrius Harris, Cordarrelle Patterson, and a host of other role players who played a few offensive snaps.

(Note: This data is pulled from the NFL Game Statistics and Information System, which includes sacks and QB scrambles as passing plays.)

A few thoughts:

David Montgomery had pretty even splits when he was on and off the field. Therefore I won’t look at him any further when I split the sample into different personnel packages below.
This is a change from 2019, when Montgomery’s presence on the field made a run much more likely, and is almost certainly due to Tarik Cohen’s injury. In Cohen’s limited 76 plays before getting hurt, the Bears only ran it 29% of the time. He clearly had the passing downs role, and Montgomery absorbed that when Cohen got hurt.
Everybody else has fairly significant changes in how frequently the offense runs when they are on the field vs. off of it, which warrants further exploration.

Different Personnel Groupings

I was curious how much the personnel groupings might influence these splits, so I looked at how frequently the Bears run the ball in different groupings. Generally, there are five skill position guys (WR, TE, RB) on the field for a given play, so I split the sample up by how many of them were wide receivers.

The more WR the Bears have on the field, the more likely they are to pass. That makes sense, but the significant difference in run frequency here means we’re going to have to look at each of these groups individually to see how players really impact the run/pass ratio when they are on the field.

3+ Wide Receivers

Let’s start with plays featuring 3 or more WRs, which means there are 2 total TE + RB. The most common setup here was 11 personnel, which features 1 RB, 1 TE, and 3 WR.

Here we see some clear trends emerge. A few thoughts:

Darnell Mooney was used pretty evenly, which is good to see. The Bears need more guys like that to avoid making their offense predictable. Mooney, Montgomery, and Robinson have now all shown they can be on the field in any situation.
Looking at the other WRs, Anthony Miller and Javon Wims have reverse splits. The Bears are far more likely to run when Wims is on the field than off it, and prefer to throw when Miller is on the field.
Notice also that virtually all of Miller’s snaps come in 3 WR sets (584 snaps total, 548 with 3+ WR on the field), which makes sense for somebody who isn’t trusted to run block.
At tight end, Jimmy Graham seems to be favored in passing plays, while Cole Kmet is favored on run plays. Kmet’s results might be skewed by early in the season, as he became much more involved in the passing game towards the end of the year. In the last 6 games, Kmet played 87% of the available snaps, meaning he was on the field for pretty much everything and would have been ignored in this study like Robinson was. Unfortunately, this data is only available for the full season, so I can’t say for sure, but this is one trend I won’t be surprised to see change in 2021.

2 Wide Receivers

Let’s look at the same data for plays with 2 WR on the field, which means there were 3 total RB + TE. The most common setup here was 12 personnel, which has 1 RB, 2 TE, and 2 WR.

A few thoughts:

Overall, we mostly see the same trends continue. Wims and Kmet skew towards the run, Miller and Graham favor pass.
The sample size for Miller is so small it can effectively be ignored. Generally, the Bears don’t like playing Miller in 2 WR sets, and they really don’t like running the ball in those situations. Miller played almost 600 snaps last year, and only 5 of them were runs with 2 (or fewer) WRs on the field.
A new trend we see here is that Mooney starts to skew towards the pass a little bit. That was not the case in 3 WR groupings, which makes me suspect it is because he gets pulled off the field for Wims (leaving Robinson in beside him) when the Bears want to run.
- Mooney was on the field for all but 86 of Chicago’s 2 WR plays, and Wims was in for 54 of those he missed (compared to them only sharing the field on 24 snaps here).
I haven’t put Cohen in any of these because he only played 76 snaps before going down with an injury, but he heavily favors pass in both 3 WR (25% run when on the field) and 2 WR (36% run when on the field) looks. That’s consistent with past years.

0/1 Wide Receivers

Finally, let’s look at plays with 0 or 1 WR on the field. This means there were 4-5 RB and TE (sometimes a 6th offensive lineman lining up at TE), so these are jumbo sets that you would expect to skew heavily towards the run.

A few thoughts:

I didn’t put run/pass splits in for Mooney or Miller because they weren’t on the field enough to make them meaningful, and Kmet because he was on the field for all but 6 of these snaps.
Looking at the 2 rotational guys, Wims heavily leans towards a run and Graham towards a pass. Shocking, I know.
The Bears really like Wims as a run blocker. He played almost as many snaps in 1 WR looks as Allen Robinson did (44), despite the fact that Robinson played more than 3x as many snaps as him on the year.

Grouping Effectiveness

Finally, I want to take a quick look at how effective the Bears were in each type of personnel grouping.

Since some of these groupings will be situational (lots of WRs on 3rd and long, lots of RB/TE and 3rd and short), yards/play might be a little misleading, so I’m also pulling in success rate from Sharp Football. This accounts for down and distance, so that a 1 yard gain on 3rd and 1 counts as a successful play, while a 6 yard gain on 3rd and 10 does not. (A full definition of success rate can be found here, if you’re curious.)

A few thoughts:

The Bears are better at passing when teams expect run, and better at running when teams expect pass. This can be seen in both yards/play and success rate.
Why, then, do they run it the most often in setups that are the least successful, and vice-versa? This reflects a stubbornness to do what you want to do, even if it doesn’t work, instead of adapting to what you’re able to do well. I would very much like to see the run/pass ratios better reflect how successful they are running and passing out of each setup in 2021.
Chicago was really good passing out of 12 formation (2 WR) in 2020. That’s not a tiny sample size either, as they threw it more than 140 times out of this look. With Kmet and Graham both back in the fold for 2021, I’d like to see them utilize this more to see if the passing success can continue.

Lessons Learned

Chicago has 3 players who have proven they can be used evenly in run and pass situations in Allen Robinson, Darnell Mooney, and David Montgomery.
Cole Kmet’s late-season usage in 2020 suggests he can become a 4th in 2021.
Tarik Cohen’s return from injury in 2021 might push Montgomery back to a run-heavy role, as Cohen continues to be a clear tell for passing plays.
Several players are pretty clear indicators of what the offense wants to do: Jimmy Graham and Anthony Miller are primarily used in passing situations, while Javon Wims is mainly used on running plays.
In general, the offense is pretty bad when they do what the defense is expecting, and much better when they are able to catch the defense off-guard. Accordingly, they would do well to run more out of 3 WR looks and pass more out of bigger sets.