Hokepoints: Who Needs SAM?

Submitted by Seth on February 25th, 2014 at 11:10 AM

Upchurch - 8173043148_1c2001efcd_o

Say uaaaahhh [Upchurch] 

Last week when I was talking about the position moves—Jake Ryan to middle linebacker, Roy Manning to cornerbacks coach, etc.—I was mostly positive in the analysis portion, explaining the move as a reaction to having their best defensive player at a defensive role that's quickly becoming as defunct as the spinning fullback.*

In the podcast Brian and Ace expressed some heebies and jeebies over the moves. I can't speak to all of those worries; who knows whether Jake Ryan can read run/pass, or if maybe Desmond Morgan's pass defense was a gaping hole the coaches were covering up in other ways. I can't even give a full answer since Brian didn't do defensive UFRs for Michigan's last three games. But I thought we might use the data we have to see whether the strongside linebacker position in Michigan's defense has been phasing out.


Spread level: rising. The vagaries of year-to-year scheduling and missing UFRs may throw off the data but Michigan's opponents indeed have been throwing out more wide receivers in their base sets as of late.

Average WRs in Formation by Situation**
Year Normal Long/2 min Short/2pt Total
2008 2.93 3.34 1.72 2.92
2009 2.55 3.14 1.83 2.59
2010 2.68 3.41 1.69 2.72
2011 2.68 3.31 1.41 2.69
2012 2.68 3.39 2.06 2.75
2013 2.81 3.31 2.17 2.87
Total 2.72 3.32 1.81 2.76

2008 was thrown off by teams going uber-spread: Minnesota, Northwestern, Utah, Illinois, and Miami (NTM) all averaged more than three wide receivers on normal downs, the former three going 4-wide more often than not. That's not too surprising given that defense had a plausible 4-3 run-stopping depth chart, but a huge dropoff if you could mitigate the DL and get past Warren and Trent on the CB depth chart. After that things normalized to a spread-leaning mix of 2- and 3-wide sets until last year.

I wish I had complete numbers. I can tell you that next year Michigan replaces CMU, UConn, Akron, Nebraska, and Iowa with Appalachian State, Utah, Miami (NTM), Maryland, and Rutgers. I can use 2013 stats (from cfbstats) to show you the playcalling breakdown of these offenses:

[If you jump first]

2013 Opponents   2014 Opponents
Opponent Pass Rush(RB) QB Run   Opponent Pass Rush(RB) QB Run
Nebraska 41% 46% 13%   Miami (NTM) 46% 13% 41%
Akron 53% 40% 7%   Maryland 50% 33% 17%
Iowa 42% 51% 7%   Utah 48% 42% 10%
Connecticut 58% 38% 4%   Rutgers 53% 41% 6%
CMU 50% 47% 3%   App State ? ? ?

Sacks counted as passing plays above. Of the teams Michigan is replacing on their schedule, only Nebraska made runs with their quarterback a significant part of their offense. In contrast Miami (NTM) is the quarterback-runningest team in FBS, and Maryland and Utah (and Appalachian State) run spread 'n shred offenses akin to Ohio State's.

That doesn't necessarily mean you get rid of the SAM, since holding the edge is a huge deal when stopping an option offense. But the point of the spread option game is to take the playside linebacker out of the running game without having to block him, either by making him the player optioned, or putting a receiver out of the box and thus forcing the SAM to cover him out there (the safety to that side is then the run defender).

This, at least, is my understanding for why Michigan's coaches are trying to get Jake Ryan into the middle of the defense and, presumably, go with a more safety-like object in his former spot. The idea is to put your best guy at getting off blocks at a position that actually takes on blocks.


Opponent DL LBs DBs
Central Michigan 3.49 3.06 4.45
Notre Dame 3.41 3.09 4.50
Akron 3.37 3.06 4.57
Connecticut 3.59 3.05 4.36
Minnesota 4.00 3.00 4.00
Penn State 3.97 2.83 4.20
Indiana 3.91 2.17 4.91
Michigan State 3.90 3.30 3.80
Nebraska 4.02 2.98 4.00
Northwestern 3.93 2.92 4.15
Total 3.76 2.94 4.30

Nickel and Diming. Many times last season Michigan didn't have a SAM in the strongside/slot area at all. Rather they went to a nickel with 3-3-5 personnel in which the nominal SAM was actually the weakside defensive end, right down to having his hand in the dirt. Michigan's avg personnel (all charted plays) by game are at right. Most games they spent at least half of the time with a fifth defensive back out there.

Ryan spent much of the year coming back from injury and wasn't 100% himself; in 2012 he was an effective edge rusher with his hand down. Nickel DE isn't a bad spot for your best defensive player, since nothing can disrupt a pass play like pressure. From recollection though our nickel defense gave  up a lot of first downs right before the pressure got there.

To get more scientific with this I broke out long (3rd/4th down 6 or more) situations and two minute drill plays, broke those into "Win" (got less than 6 yards) and "Loss" (gave up 6 or more yards), and looked at which players Brian tagged as the key to that play. W-L results:

Position 2011 2012 2013
Linemen 20-8 15-7 29-5
Cornerbacks 16-10 19-14 17-20
Safeties 13-9 9-10 7-11
Middle LBs 5-6 13-13 15-13
SAMs 2-2 15-2 5-3
RPS Stuff 8-7 5-2 12-4
Totals 64-42 76-48 85-56

Of course DL are going to chart better—a loss for them is giving up contain, and their pressure metric isn't counted. Corners left 1-on-1 are going to be targeted. I think that 15 and 13 record for the middle linebackers is pretty ugly both in results and frequency, though it's no different than the previous year with Demens. Michigan's success rate on 3rd and long and two-minute drills wasn't any different (about 60% all three years). Most notable actually is how little the SAMs figured into it versus the year before, when Jake Ryan was a high-frequency contributor.

Conclusion: Michigan probably needs more than one Jake Ryan.


* [This position keyed the Mad Magicians offense back in the 1940s. Think of it as 35% quarterback, 35% blocking fullback, and 30% running back. Jack Weisenberger was born to play spinning fullback in Fritz Crisler's offense like Jake Ryan is made to be a SAM in a 4-3 under, but if you have Weisenberger on your team in 2013 you don't make him a fullback.]

** ["Long" = 3rd or 4th down and >5. "Short" = 3rd or 4th down and <3 or any down inside the 3 yard line. "2 Min" = situations where the offense was obviously in hurry-up mode at the end of a half.



February 25th, 2014 at 11:27 AM ^

A couple points:

1)  Why would you say the linebacker numbers are ugly when they have a better success percentage than both the corners and safeties?

2)  Perhaps the lack of impact from the SAM was due to the position being filled by a back up for most of the year.  A full season of Jake Ryan in 3rd/4th and long, and there would likely have been more of a positive impact from the position.


February 25th, 2014 at 11:33 AM ^

You can put me in the "bit nervous about all these moves" camp too.  Year four of the Hoke regime was when the stability of the program and the fruition of Hoke's recruits should be taking hold IMO and now we've musical chairs on defense other than Mattison and our best defensive player gets to learn a new position.  

That doesnt exactly inspire unbridled confidence in for the upcomng year.

STW P. Brabbs

February 25th, 2014 at 12:21 PM ^

I think Mattison and Hoke want to have an aggressive, dominating defense, and they're going all in to make that happen.  Especially because the prospectus of the OL next year is dark and full of terrors, I don't think standing pat with a bend-but-don't-break type of outfit is a good idea, and I'm glad things are being shaken up.

If the defense was a bit of a tire fire last year and these moves were being made, I'd be more anxious about them.  They are, instead, being made from a position of relative stability and strength, so to me there's little sense of flailing.


February 25th, 2014 at 11:38 AM ^

I don't think putting Ryan in the middle is a bad thing. Putting Ryan there would probably help with run D. When he is lined up to one side or the other, teams can simply run away from him and effectively neutralize him. Now that he is in the middle he can get to both sides of the field faster.


Jeff M

February 25th, 2014 at 11:46 AM ^

I can buy Ryan at MLB as an upgrade, particularly against these spread-y teams. The major concern is the gaping hole that leaves at SAM. Even if the "SAM" is really the Nickel DE for ~60% of downs (which Beyer could hld down), for the other ~40% of downs that's a huge hole on the edge. 

Maybe I'm interpreting this wrong -- maybe they go Ross/Morgan/Ryan against traditional teams/in run downs and Ryan is a "Nickel MLB." That's still asking a lot of Ryan, and lot of whoever replaces him on the edge.


February 25th, 2014 at 11:51 AM ^

Don't care. But we need pass rushers. Will never win the big ten it we don't have a consistent pass rush. We can't go against teams like Akron Indiana PSU and NW and struggle to get to the QB.


February 25th, 2014 at 1:26 PM ^

It seems like the coaches are throwing everything they have at maximizing the production of the LB crew both with the coaching changes and the position changes.  But the reality is that the true problems were with the DL.  We couldn't consistently get pressure and we couldn't keep the LB free.  So now we're tinkering with the deepest and most talented position on the field to cover up for the weakest?  Seems.....odd.  Maybe they see the writing on the wall that help is NOT on the way at DL and they need to try more exotic options.  Let's face it - unless we see a huge improvement from someone unexpected on the DL we're treading water there.  The other option is that Ross will end up being a dynamic player on the edge where his speed and quickness are more effective and Ryan can blow up releasing guards and still make the tackle.  Let's hope......  


February 25th, 2014 at 5:53 PM ^

You play to your strengths, not to your desires.  Mattison is showing the flexibility Borges didn't have, albeit maybe a full season too late.

In terms of concepts "4-3" is generally associated with a disrupting DL and clean-up LBs wheras 3-4 is a "two-gap" system where the DL eats space and the LBs make plays.  I want to say that's "outdated", but it's really inaccurate to begin with.  It's better to go with Space Coyote's concepts of "inside out" and "outside in".  Mattison's preferred mode is "outside in", where the DEs and SLB hold the edge and DTs get push to squeeze the play until there's nowhere to go.  But that requires stout DTs and playmaking DEs and Michigan currently has neither.  If blockers are getting to the second level, the execution has already failed.  3-4 is typically known for having a huge space-eating DL, but this is also not always accurate IMHO.  A 3-4 can also be deployed with swarming blitz packages that don't expect the DL to do it all themselves.  They shoot the gaps and force the ball outside, where the corners and OLBs chase.  That's "inside out", and Michigan has the personnel to do that.  Again -- the 4-3 vs. 3-4 isn't important; the key here is where the disruption is going to come from.  In 2014, it's going to be the linebackers.

The 2011 defense was very good because all the tools GERG badly misused were in place -- Mattison wanted the DL making plays and RVB and Mike Martin were up to the task.  The linebackers were a weakness then.  In 2013, the linebackers were deep and the DL  thin.  Pass rush and run stuff just weren't coming from the DL and that looks to be even more lopsided next season.  In that situation, sticking to a concept where the DL has to make plays and your best player just contains one side is madness.

The 2014 defense has quick CBs and LBs, so it makes more sense to put your block-beating linebacker on the inside, and a bunch of killer bees like Ross and Peppers to the outside.  The pre-snap reads and formations won't change dramatically from the 4-3 under; this is really just a personnel adjustment vs. spread teams that option off JMFR.  Instead of a slot receiver drawing our best pass rusher away from the play while the LBs eat blocks, JMFR can beat blocks on the inside and flush the ball toward guys like Countess, Ross and Peppers.  Against sets featuring a TE, we can always move him back out to SAM.

This is just how I see it, anyway.  That said, I have no idea why they're shuffling the coaches so dramatically.

blue in dc

February 25th, 2014 at 7:36 PM ^

If you start with the idea that you don't want Hoke to be a position coach, you have to do something with the defensive line. If you don't want Mattison coaching four positions and you have a guy with many years of experience coaching d-line, you think, I'll move Smith to defensive line. You also think, I was planning on doing some things differently with my linebackers to become more aggressive and my d-coordinator was actually an NFL linebacker coach, moving him to linebacker could make sense.

Now comes my one challenge, one of my best recruiters was already coaching one of the linebackers, do I keep him there, or does it make more sense to move him somewhere else? If you are comfortable with Mattison coaching all of the linebackers, it's a waste of a coaching position to have him stay coaching a single linebacker position.

The one defensive position you are now uncomfortable with is Dbacks where you have one coach, coaching five positions, but Manning doesn't have experience with d-backs. Should I fire him and hire a second experienced d-backs coach? Well, I really like the job he's been doing recruiting, so I don't want to get rid of him. Also, the guy who used to coach all five d-back positions is still here and could act as a mentor. There's a bit of risk, but big picture, when I factor in both day to day coaching and recruiting, I am comfortable taking that risk.

Other than cb's, nothing drastic and all very logically. At cb's, a calculated risk, but Manning isn't an unknown quantity, so Hoke/Mattison are taking that risk with much more knowledge at there disposal.