Schneider's Strategem: The Running Back Draft

According to John Schneider, the 2016 draft class is the deepest talent pool in the last seven years. While some few critics may disagree with Schneider's assessment of the draft pool itself, the Seahawks' General Manager is globally recognized as the authority on the Opinions of John Schneider.

Which is enough to make things interesting for Seahawk fans. With that in mind as the big picture, I've been looking over Seattle's recent draft haul for the intricate subtleties by which Schneider is, possibly, attempting to maximize the value of this once-in-a-Tribulatiennial draft class. In fact, I noticed so many tidbits that I decided to start with an article focusing solely on the running backs.

Seattle Drafts Three Running Backs

That sounds like a lot when you consider that the rest of the league combined only drafted 17. But it sounds like even more when you say that the Seahawks drafted 447% more running backs than the average non-Seahawk team.

So what's up with that? Conventional wisdom holds that the importance of the running game has declined, and teams are less likely to expect big gaps in talent and value at the running back position. This made me curious about the number of players drafted compared to the number signed as undrafted free agents. A higher number of UDFA's would be expected if teams really thought the position fungible.

Using SB Nation's tracker for 2016 UDFA's, I compiled a table for your viewing pleasure, as well as a confirmation of my identity for regular Field Gulls readers (I like tables). :

(The UDFA tracker and Pro-Football-Reference.com's draft information have different listing standards involving the frequency of specific positions, such as "FS" for Free Safety and generic positions such as "DB" for defensive back. So comparisons of defensive backs, offensive linemen, and defensive linemen should use the totals at the bottom. Also, where a UDFA had two positions listed, such as "OLB/DE", the player was counted half for each.)

Position	UDFA 2016	UDFA%	Drafted 2016	Draft% 2016	Drafted 2015	Draft% 2015
Guard	10	2.2%	17	6.7%	18	7.0%
Tackle	27	5.9%	18	7.1%	20	7.8%
OL	31	6.8%	6	2.4%	8	3.1%
RB	23.5	5.1%	20	7.9%	18	7.0%
FB	12	2.6%	3	1.2%	4	1.6%
DT	28	6.1%	20	7.9%	16	6.3%
DE	22.5	4.9%	19	7.5%	21	8.2%
DL	30	6.6%	3	1.2%	6	2.3%
CB	40	8.7%	31	12.3%	29	11.3%
S	22	4.8%	19	7.5%	13	5.1%
DB	28	6.1%	1	0.4%	4	1.6%
LS	12	2.6%	1	0.4%	1	0.4%
K	12	2.6%	1	0.4%	0	0.0%
P	4	0.9%	3	1.2%	1	0.4%
LB	31	6.8%	34	13.4%	36	14.1%
OL total	68	14.8%	41	16.2%	46	18.0%
QB	10	2.2%	15	5.9%	7	2.7%
RB/FB tot	35.5	7.8%	23	9.1%	22	8.6%
WR	77.5	16.9%	31	12.3%	34	13.3%
TE	37.5	8.2%	11	4.3%	20	7.8%
ST tot	28	6.1%	5	2.0%	2	0.8%
DL tot	80.5	17.6%	42	16.6%	43	16.8%
DB tot	90	19.7%	51	20.2%	46	18.0%
ALL	458	100.0%	253	100.0%	256	100.0%

Looking only at the number of players taken, this is not quite what we expected in comparison to other offensive skill positions. Teams drafted a total of 42 TE's and WR's, while picking up 115 UDFA's; that's 174% more UDFA players at the receiving positions. But they drafted 20 RB's and added 23 1/2 UDFA's, an increase of just 17.5%. Maybe there weren't that many good running backs left over, after all.

Draft Picks vs UDFA's Production

Or... maybe we need to look at what sort of production actually comes out of drafted players and free agents in future years. Using the excellent search engines on PFR, I searched 2014-2015 for all wide receiver seasons accruing at least 700 receiving yards, and all running back seasons accruing at least 500 rushing yards. The following scatter plots show draft position on the x-axis and annual yardage on the y-axis.

(Many players showed up for both seasons, in which case they are represented as two different data points. Through extrapolation, pick #295 was estimated as the draft position of UDFA's for the purpose of calculating the correlation and slope.)

This tells us nothing about success rates, of course, because the data completely ignores players who didn't meet minimum production. But put yourself in the mind of a GM looking around the league, and this is probably the data-- or at least, the general impression-- that you are working with. It precisely answers the question "where do productive players come from"?

And it accords with the conventional wisdom. Fully 79% of the qualifying wide reciever seasons came out of the top three rounds of the draft, compared to just 55% of the running back seasons. At the other end, only 6 WR seasons came from UDFA's (6%) compared to 16 RB seasons (21%).

What the hell, right? It looks like UDFA's work fine at running back, and you need to spend high draft capital to get productive receivers. But from our previous table, 46% of rookie running backs are draftees, and only 29% of rookie receivers are draftees (the rest being UDFA's). Teams are getting their running backs out of the draft, but stocking up on undrafted receivers. Everything is backwards.

So: I'm going to propose a hypothesis and see where it leads us. Let us assume that, for the number of rookies coming to training camp, the top-to-bottom talent dropoff for wide receivers and running backs is exactly the same. But the league's General Managers tend to believe that there is a steeper dropoff at wide receiver.

If true, it would produce the same results: Most productive wide receivers are highly drafted simply because teams load up on wide receivers in the top rounds, the effect being a matter of numbers rather than talent. And more productive running backs come from the UDFA class and latter rounds simply because there are still talented backs available when teams don't draft enough.

Also if true, we would expect a weaker correlation between draft position and production for receivers (because teams are wasting high picks), and a stronger correlation between draft position and production for backs (because the talent drop-off is real, and teams who spend draft capital get significantly better players).

Well, then: The correlation coefficient for receivers is -0.138. For running backs, it is -0.227 (don't worry about the minus sign, it's magnitude that counts). Zing!

I think that Seattle's 5th-round selection Alex Collins is a player that half a dozen other teams expected to pick up in the 6th round, each believing that there would be little interest from others; and 7th-round pick Zac Brooks was similarly expected to be available as an undrafted free agent. But Schneider beat them to the punch, got good value for relatively low picks, and left the other teams out in the cold. Based on SB Nation's latest, only 15 teams picked up any UDFA running back. I'm sure they all have enough bodies in camp to go through the motions of competition, but once the regular season starts they aren't going to have the talent or depth the Seahawks enjoy-- at what is, incidentally, far and away the most injury prone position.

C.J. Prosise: A Perfect Fit?

The Seahawks grabbed their first running back in round 3, making Notre Dame's C.J. Prosise the fourth running back drafted overall. Prosise is a converted receiver. The Seahawks will throw him the football. Also, Russell Wilson is short and Kam Chancellor is painful.

That's the obvious stuff out of the way. But my first thought during the draft was that Prosise is the ideal player to have while Jimmy Graham and Thomas Rawls are recovering from injury. He fills out the depth chart at running back and gives Seattle another legitimate downfield route-runner. In other words, Prosise is going to save us a roster spot. Instead of carrying both an extra back and receiver, Prosise fills both roles, which will allow the Seahawks to retain one more lineman or defensive back or what-have-you from this talented rookie pool.

I've got a pretty good idea about how he'll be used as a receiver, too. Russell Wilson loves throwing from an empty backfield. I personally dislike the formation as a matter of principle, because it frees the opposition defense from having to guess the play type. However, Wilson is a pretty good threat to run himself; and, more importantly, it's hard to argue with results:

Offensive Team Efficiency, 2012-2015 (Football Outsiders DVOA)

TEAM	4-year DVOA average	Total DVOA rank	4-year Weighted DVOA Average	Total Weighted Rank	4-year Pass Average	Pass Rank	4-year Rush Average	Total Rush Rank
NE	19.0%	1	20.6%	2	38.2%	1	2.3%	6
DEN	16.7%	2	14.9%	3	36.8%	2	-1.3%	11
SEA	15.8%	3	20.8%	1	32.5%	3	14.8%	1
GB	13.8%	4	9.7%	5	28.5%	6	2.5%	5
NO	12.3%	5	12.1%	4	28.1%	7	-2.5%	15
PIT	10.1%	6	9.6%	6	28.7%	4	-6.8%	17
SD	5.3%	7	3.6%	8	28.6%	5	-12.4%	28
CAR	5.1%	8	5.6%	7	14.2%	13	3.3%	4
CIN	4.0%	9	1.5%	13	18.5%	8	-1.9%	13
DAL	3.7%	10	3.2%	9	15.9%	10	0.1%	9

The Seahawks have, without question, one of the league's elite offenses. Their average end-of-season ranking (weighted DVOA) is highest in the league and the average overall ranking is third. Passing offense alone (which does count all those sacks and does not include Wilson's scrambles) also ranks third-- that's better than the Saints, Steelers, and Packers.

Seattle's empty backfield saves the defense from guessing the play type based on formation, but not based on personnel. The fifth receiver is frequently a running back who initially lines up in the backfield, either as a pre-arranged deception or as an actual running option (with an audible being given based on the defensive response). And the threat to run is legitimate, because the Seahawks don't limit the formation to obvious passing downs. Three quarters of the receiving production from Marshawn Lynch and Robert Turbin came on 1st and 2nd down:

Running Back receptions on 3rd and 4th down, 2012-2014
Turbin: 14 rec, 118 yards, 0 TD
Lynch: 22 rec, 182 yards, 2 TD
Combined: 26 rec, 300 yards, 2 TD

Running Back receptions on 1st and 2nd down, 2012-2014
Turbin: 29 rec, 309 yards, 2 TD
Lynch: 74 rec, 697 yards, 5 TD
Combined: 103 rec, 706 yards, 7 TD

I also took notes from the Seahawks' 2015 regular-season finale at Arizona (it was the only game left on my TIVO). Christine Michael isn't the receiving threat that Lynch or Turbin was, but Fred Jackson is, and it was interesting to see how he lined up.

Jackson stayed in the backfield (at least up to the snap) throughout the 2-minute drill, when the game clock didn't really allow motion. He also stayed back on a handful of second half plays, when protecting Wilson was more important that actually advancing the ball (thanks to a 24-point lead).

That left seven other plays in the first half when the clock wasn't a factor and the Seahawks were still trying to score. Jackson lined up as a receiver directly out of the huddle twice. He motioned to a receiver position and returned to the backfield twice, and three times he started in backfield but was split wide at the snap.

That's how I foresee Prosise being used as a passing target-- not in place of a receiver with another running back on the field, and not limited to third down, but as an optional running threat who can motion wide. Imagine Prosise in the backfield with Doug Baldwin, Tyler Lockett, and Jimmy Graham across the line. Put Luke Willson in as the fifth position player and you have a 2-1 power running formation-- decidedly below-average in terms of run-blocking capability, to be sure, but would anyone dare to counter that with a base defense? Even a nickel package is going to be potentially over-matched, as the Seahawks will have more receiving talent than a lot of teams can get with five wide receivers.

Take out Willson and put in rookie blocking tight end Nick Vannett for a stronger running option, or Jermaine Kearse for a better passing option. I especially like Prosise as a backfield motion threat against the dreaded pass rush of a 3-4 alignment (such as that used by the Super Bowl Champion Broncos). Teams that play a 3-4 are comfortable with their nickel package against 3 wide receivers, but the threat of 4 or 5 downfield route runners will force one of their stud pass-rushing outside linebackers to leave the field or else play pass coverage against a tight end/running back, or they'll have to pull out an interior defensive lineman. The last option is preferred in obvious passing situations, but with Prosise and two tight ends it makes for a very soft front to run against.

While Prosise is certainly no Marshawn Lynch when it comes to running the ball, his dual-threat capability could make him just as effective overall in certain packages. And "just as effective as Marshawn Lynch" is pretty good for a guy who isn't even the team's #1 running back.