During the preseason I was goofing around with wide receiver targeting stats by Bill Connelly*, and Ace asked me if it says anything about anything if a team is targeting its tight ends more than its receivers. At the time it seemed Michigan was about to do that. They haven't:
But once things shake out it wouldn't be that surprising if it's Darboh and Butt then a bunch of low-amplitude dudes. The more the season has progressed, the more it seems Ian Bunting and Henry Poggi are going to siphon snaps and targets from Grant Perry and Drake Harris. Jehu Chesson gets ignored even when his cornerback has fallen down. We can compare this distribution to the rest of the Big Ten:
| School | Targets | Yards | Off S&P+ | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| WRs | TEs | RBs | WRs | TEs | RBs | S&P | Rk | |||
| Rutgers | 53% | 31% | 16% | 82% | 9% | 9% | 25.7 | 86th | ||
| Michigan | 53% | 32% | 14% | 73% | 19% | 8% | 32.4 | 48th | ||
| Northwestern | 57% | 26% | 16% | 63% | 22% | 15% | 22.4 | 105th | ||
| Iowa | 65% | 18% | 17% | 79% | 13% | 9% | 33.1 | 44th | ||
| Wisconsin | 65% | 13% | 21% | 47% | 40% | 13% | 33.4 | 42nd | ||
| Maryland | 68% | 15% | 17% | 79% | 18% | 3% | 28.6 | 73rd | ||
| Purdue | 70% | 8% | 22% | 75% | 13% | 12% | 27.3 | 80th | ||
| Indiana | 72% | 18% | 10% | 87% | 3% | 11% | 36.9 | 25th | ||
| Minnesota | 73% | 13% | 15% | 57% | 32% | 11% | 29.5 | 67th | ||
| Penn State | 74% | 18% | 8% | 87% | 7% | 6% | 29.2 | 68th | ||
| Illinois | 77% | 12% | 11% | 79% | 13% | 8% | 24.6 | 95th | ||
| Michigan State | 79% | 17% | 4% | 76% | 7% | 17% | 38.2 | 19th | ||
| Nebraska | 80% | 9% | 10% | 62% | 23% | 16% | 36.9 | 24th | ||
| Ohio State | 83% | 8% | 9% | 67% | 16% | 17% | 35.4 | 33rd | ||
For the above I counted OSU's H-backs as receivers, fullbacks as RBs, and Northwestern's "superbacks" as tight ends. It's early in the season so there's still a ton of mess in those numbers. So lets get some more data and see what we find.
[after the jump: two blobs jousting]
Connelly's targeting stats go back to 2005 so there's 10 years of data to play with. Brian Fremeau's efficiency index (FEI), which measures drive efficiency, goes back to 2007. So I plotted every team since 2007 by how much they target tight ends versus receivers** and…
Each dot is a team, plotted by the % of targets that went to receivers and the % that went to tight ends. There is no correlation between throwing to tight ends and offensive efficiency. In fact how much you target receivers versus TEs and backs is so not correlated that Excel chose to express the r-squared of 0.00002 in scientific notation. It's the same if you do it by yards instead of targets.
The teams that threw least often to receivers were the dedicated triple-options. Those who got the greatest proportion of their yards from their TEs were fine unless hey had Sun Belt talent. A top five:
Percent of team's total passing yards by tight ends
- 50% – 2007 Wisconsin: 33rd in FEI
- 46% – 2012 Stanford: 34th in FEI
- 44% – 2014 FIU: 125th in FEI
- 43% – 2012 Florida: 43rd in FEI
- 43% – 2011 Western Kentucky: 96th in FEI
Harbaugh's last Stanford team was 6th in FEI and got 37% of its passing yards from tight ends (12th highest in the study). The list has sprinklings of teams that simply didn't have any good receivers or ultra-good tight ends. It also has Harbaugh/Wisconsin teams that played a lot of snaps in I-heavy formations (one receiver, two TEs and a fullback).
Is Michigan that TE heavy?
So far Darboh, not Butt, is Michigan's top target. And we've seen Chesson get wiiiiiide open downfield a ton. But it's also not hard to see Butt become a bigger part of this offense since a good portion of the throws Rudock hasn't made have been those where Butt has a step or two. He's obviously the offense's best player. And Bunting, even though he's gotten just four passes his way, looks like a weapon Michigan can start using more. So let's say, just for fun, that Michigan's passing offense does become 50% tight ends.
| Player | Pos | Tgts | Yds |
|---|---|---|---|
| Travis Beckum | TE | 115 | 982 |
| Garrett Graham | TE | 54 | 328 |
| Kyle Jefferson | WR | 42 | 412 |
| Luke Swan | WR | 39 | 451 |
| Paul Hubbard | WR | 26 | 305 |
| P.J. Hill | RB | 19 | 89 |
| David Gilreath | WR | 9 | 10 |
| Zach Brown | RB | 8 | 23 |
| Marcus Randle El | WR | 4 | 9 |
| Chris Pressley | FB | 3 | 8 |
| Andy Crooks | TE | 2 | 4 |
| Xavier Harris | WR | 2 | 30 |
| Daven Jones | WR | 1 | 9 |
| Sean Lewis | TE | 1 | 7 |
What Would It Look Like?
The outlier, that one red dot invading the blue, is a fine example. That is 2007 Wisconsin, the first coached by Bret Bielema with offensive coordinator Paul Chryst. They had one burly NFL draft pick receiver, 6th rounder Paul Hubbard; a Chesson-like object in sprinter Kyle Jefferson; and the Dileo-esque Luke Swan.
Quarterback was Tyler Donovan, a guy I remember for having pretty good feet and getting damaged a ton (this is long enough ago that announcers would talk about his toughness for coming back in after his third head shot, not long enough ago you shouldn't wonder what Bret Bielema scores on ethics). and the excellent Wisconsin RB du jour was P.J. Hill, and as you might expect from a Wisconsin team they ran most of the time, often from heavy sets. When they did pass Donovan leaned heavily on tight end Travis Beckum, who had 903 yards on 115 targets. Second-leading receiver was another tight end, Garrett Graham, who would lead the team in targets in 2008; if you're drawing one-to-one comparisons he's more of an Ian Bunting/Henry Poggi/Khalid Hill mashup (Graham is now with the Texans). Their distribution is at right/above.
Michigan's numbers are too messy now for most guys to project (Drake Harris won't end up with 20 targets for 26 yards), but Darboh would be 104 targets for 787 yards right now and Butt is on pace for 562 yards on 78 targets. As Rudock grows more comfortable with the offense, it's not unreasonable to expect Butt's numbers to jump up to Beckum usage. Anyway the point is don't worry if the offense is coming via TE.
-----------------
* [The play-by-play data the teams provide have lines like "Jake Rudock pass incomplete to Amara Darbouh" [sic] that he cross-references with rosters and such.]
** [I had to reclassify a lot of guys listed as other things on their rosters. Things like "FL" were simple enough, but "SB" could be a fullback (Navy), tight end (Northwestern) or running back (Rich Rod teams). QBs with more than 15 targets I counted as receivers; the rest I discounted along with OL and other obvious trick play dudes. For LBs and SSs etc. with more than 15 targets I looked up what they were used as; the rest I removed.