Now that we’ve seen Chicago’s new offense play four games, it’s time to examine what exactly it looks like. We’ve seen them run 271 plays, and while that’s still a fairly small sample size, it’s big enough that we can begin to pick up trends, search for predictable patterns that opposing defenses might begin to pick up on, and see if there are any situations their current approach could be improved.
Today we’re going to focus on running backs Jordan Howard and Tarik Cohen, examining how much they’re playing, how effective they’ve been, and how they’re being utilized.
Snap Counts and Efficiency
First I want to look at how frequently each running back is playing, and how their presence on the field impacts the offense’s performance. Data is from The Quant Edge.
(Note: I’m using success rate here instead of yards per play. That is to account for down and distance context. A two-yard play on 1st and 10 is bad, while a two-yard play on 3rd and 1 is good. The general idea is that a successful play keeps you ahead of the chains, but an exact definition is available here if you’re curious.)
A few thoughts:
- Howard is actually playing typical lead RB snaps for an Andy Reid offense. As I noted this offseason, Kareem Hunt played 65% of the snaps in Kansas City last year. This is in stark contrast to the Philadelphia Eagles style-rotation I thought would make more sense. It’s worth noting the 2 split snaps almost exactly 50/50 in the Tampa Bay game. I wonder if that’s more what we’ll see going forward.
- The run/pass splits for when both of these players are in vs. out of the game are too lopsided. A 30% swing when Howard exits the game and 20% swing when Cohen enters the game should not be the case. This is too predictable and makes it too easy on the defense.
- Some of these numbers are related, since Howard and Cohen basically swap being on the field. They’ve only shared the field on 23 snaps so far this year, so the run game being more effective with Howard off the field is the same as saying the run game gets more effective when Cohen is on the field. Again, I think this might have more to do with defenses gearing up to stop the run when Howard is in the game and not expecting it when Cohen is in.
Where They’re Running
Another thing I looked at this offseason that I want to return to now, is where each running back is getting their carries. Specifically, I’m interested in how often they’re carrying the ball inside of the tackles vs. outside of the tackles, and how effective they are in each situation. That data (compiled from Pro Football Reference’s Game Play Finder) can be seen in the table below (YPC = yards per carry).
A few thoughts:
- This is also following more of a KC model than a PHI one; the Chiefs largely run inside while the Eagles ran a lot more outside runs despite running ostensibly the same offense. It appears Nagy has largely implemented a KC-style Reid offense, though he is more Philly-like in how little he uses a fullback (Michael Burton has played 6% of offensive snaps, while KC’s fullback is usually around 20% and Philadelphia doesn’t have a fullback).
- Jordan Howard has really struggled running the ball inside and out this year, but he’s been especially awful outside of the tackles. Being better inside than out matches his performance in 2017, while his breakout 2016 season saw him more effective outside than in. More on Howard’s struggles in a minute.
- Tarik Cohen really struggled running the ball inside last year, but he’s done it fairly effectively so far in 2018. That’s good news for the Bears, as it has allowed them to get a nice balance between inside and outside runs with Cohen on the field.
I now want to take a minute to focus individually on each player, starting with Tarik Cohen. This past offseason I looked at how the Eagles utilized Darren Sproles as a blueprint for how Tarik Cohen should be used in this offense. Let’s check in and see how well that comparison is stacking up.
On the surface, that comparison seems to be reasonable. They are similar in touches per game and carry/catch ratio, but Cohen is running outside less frequently than Sproles did. Given Cohen’s small stature, it might be wise to have him run outside a little bit more often to try and avoid big hits.
Here’s where a small sample size can become an issue though. The Tampa Bay game really skews the numbers, since Cohen had 20 touches in that game and 21 in the first three weeks combined. Cohen was a huge weapon against Tampa Bay, racking up 53 yards on the ground and 121 yards (and a TD) through the air. His role in the offense shot through the roof, and the Bears’ offense worked much better. Is that because of Cohen, or because of other factors (like running more often out of shotgun, or TB’s defense being really bad)?
Only time will tell, but the Bears can’t expect Cohen to hold up if he’s getting 20 touches a game.
Side note: Cohen has been a real weapon as a punt returner so far this year. That’s been valuable as well.
Now let’s move to Jordan Howard, who is having the worst season of his career so far. I want to take a minute to look at what’s going on in the run game in a few different ways before also looking at his role in the passing game, which has been a pleasant surprise.
First I want to look at how far his carries are going to see what’s accounting for his drop in yards per carry. I used the Pro Football Reference Game Play Finder to sort each carry in his career so far and compile the data below.
When I look at this, I see two changes taking place. First, 5% of his runs are going from short gains of 1-3 yards to stuffs at or behind the line of scrimmage. Second, 6% of his carries are going from big runs of 11+ yards to medium gains of 4-6 yards.
It’s worth noting that’s a small sample size; Howard has 64 runs so far this year, so 5% is only 3 runs and 6% is 4 runs. So he’s literally just a few plays away from being where he’s normally at. For the stuffs behind the line, I think that’s on the OL and play calling. Howard takes some time to get up to speed, and so the blockers need to keep him clean at least through the line of scrimmage. For the big runs, that’s on Howard to break a tackle in the open field. He hasn’t done that as well as past years so far.
Next I want to look at what types of defenses Howard is running against, using Player Profiler to provide the data. I noted this offseason that running backs in both Philadelphia and Kansas City saw more than half of their carries against 6 man boxes in 2017, and I expected Howard to excel in the run game in this offense because he averaged over 6 yards per carry against those looks in his first two seasons. As you’ll see in the table below, that hasn’t happened.
To me, this table says two things.
- The new offense is not doing its job and getting Howard carries against light boxes. He should be seeing close to half of his carries against nickel fronts, and very few 8 man fronts.
- Howard is still struggling to run the ball regardless of the situation. He’s doing better against lighter fronts, but worse against each front than he did in 2017. I don’t know if that’s on Howard, the offensive line, or play calling where the defenses are expecting run, but the Bears need to figure that out in a hurry.
I also want to note that, according to Sharp Football, Howard might not actually be struggling at all. He tracks success rate for RBs based on different situations they run the ball and notes that Howard has a success rate that is 9% higher than expected so far this year. Again, success rate accounts for down and distance when evaluating the outcome of a play (full definition here). To be honest, I’m not entirely sure how accurate this is, as a similar measure on the same page gives Howard a +5% compared to expected success rate. I just thought it was worth pointing out as something that suggests Howard’s low yards per carry might be due more to the situation in which he’s carrying the ball than anything else.
Football Outsiders also has Howard not really struggling. Through 4 weeks, he ranks 10th in the NFL (out of 38 running backs) with a 52% success rate. So maybe it is just as simple as Howard’s not breaking big runs right now, which drops his yards per carry, but he’s still getting decent yardage a good percent of the time.
Lastly, I want to look at Howard’s usage in the passing game, which is one area where he’s actually done quite well in 2018. I said this offseason that Howard needed to be more involved and more efficient in the passing game if he wanted to keep his role as an every-down back, and so far he’s done just that.
Every down backs typically see 3-5 passing targets per game, catch 75-80% of their targets, drop less than 10% of their targets, and average 6-8 yards per target. Howard was nowhere near any of those metrics in 2017, but he’s hitting every single one of them so far this year. Through four games, he’s averaging 3.0 targets per game, catching 83% of his targets, averaging 6.5 yards per target, and he hasn’t dropped a pass. It’s a small sample size to be sure, but bodes very well for Howard’s future in the NFL if he can continue it.
Finally, I want to end with a brief look at when Howard and Cohen have been on the field together. As I noted above, this has not happened very often this year; they have only shared the field for 23 plays. This is too small of a sample size to draw any definitive conclusions from, but in looking at those 23 plays, I see enough to make me want the Bears to do this more (all stats from the NFL Game Statistics and Information System):
- 13 passes, 124 yards (9.5 yards per attempt)
- 10 runs, 53 yards (5.3 yards per attempt)
- 3 plays of 15+ yards
Jordan Howard and Tarik Cohen are two of the best players on Chicago’s offense. Just because they ostensibly play the same position doesn’t mean they can’t share the field. They both stress defenses in different ways, and the Bears need to figure out how to use that to their advantage with both of them on the field challenging defenses together more regularly.