Writers of Pro Football Prospectus 2008

printable version

Football Outsiders

Innovative Statistics, Intelligent Analysis

Need Help with KUBIAK or Premium?

blog advertising is good for you


16 May 2011

RB vs. Different Defensive Schemes

by Aaron Schatz

I was intrigued by this John Clayton story from Friday in which Jim Schwartz suggests that the Lions drafted Mikel Leshoure because they want to use different kinds of running backs against different kinds of defenses. "You play 3-4 teams that are two-gapping," says Schwartz, "You need a big back who can run through some arm tackles. You want to get guys matched up on different teams, you need guys who can match up and beat linebackers and people that want to play man and trick coverage up for a certain player."

Clayton extrapolates this idea to a more general point: That smaller running backs are struggling with the rise of 3-4 defenses around the league. Is this true? I decided to give it a look -- albeit a quick, dirty, imperfect look.

I took all the running backs in 2010 who gained at least 750 yards on the ground and split them into two groups: larger backs and smaller backs. The split took place at 30.3 BMI, although I split the two backs with 30.3 BMI: Knowshon Moreno, at 5-foot-11 and 217 pounds, went into the "larger backs" category, while Ray Rice, at 5-foot-8 and 199 pounds, went into the "smaller backs" category.

I then looked at each back's performance against both 3-4 and 4-3 defenses. (I didn't count games against Buffalo, which switched back and forth and couldn't quite figure out what it was doing on defense last year.) I looked at yards per carry and DVOA, but also at broken tackles, because Clayton specifically noted the idea that you need a larger back to beat 3-4 defenses because they need to be able to break arm tackles. Our broken tackles numbers represent a number of different plays -- a broken arm tackle, a broken ankle tackle, juking the defender out of his shoes -- but I figured the results might be interesting.

Here are the averages for each group against the 3-4 and then against the 4-3.

LARGER BACKS DVOA Att Yds Avg BT BT/Att
vs. 3-4 4.0% 96 402 4.17 7 0.07
vs. 4-3 9.8% 122 511 4.18 12 0.10
SMALLER BACKS DVOA Att Yds Avg BT BT/Att
vs. 3-4 4.5% 112 487 4.36 9 0.08
vs. 4-3 9.3% 138 658 4.76 13 0.09

Looking at yards per carry, it doesn't look like larger backs do better than smaller backs against 3-4. But it does look like an advantage that smaller backs have compared to larger backs against 4-3 tends to be neutralized against 3-4. However, the DVOA ratings are basically the same for both sets of backs against 3-4 and 4-3, which suggests that the phenomenon Clayton describes may not exist. And broken tackle numbers suggest that, on average, smaller backs actually break more tackles against 3-4 defenses than larger backs do.

Now, like I said, this is a quick-and-dirty study. A look at 25 backs in one season doesn't really disprove Clayton's thesis about different types of backs against different types of defenses. But the phenomenon certainly doesn't stand out with one quick look at the stats. Ironically, one back who really doesn't adhere to this thesis is the back Clayton uses as his biggest example: Adrian Peterson. Here is how Peterson performed against each type of defensive front in 2010:

ADRIAN PETERSON DVOA Att Yds Avg BT BT/Att
vs. 3-4 15.5% 159 715 4.50 14 0.09
vs. 4-3 -5.7% 109 473 4.34 10 0.09

Peterson didn't seem to have too much problem with the Packers defense last year -- 203 yards on 42 attempts -- and had big days against Arizona and Miami. He had his worst running day of the year against the 4-3 Giants. Clayton says that Peterson is the league's best at breaking arm tackles, and that he's struggled against 3-4 defenses in recent seasons. The two ideas don't seem to go together, and last year's numbers suggest that the former is more true than the latter.

I should note that I'm clearly working with some old player measurements that need to be updated -- for example, my files have Ray Rice listed at 199 pounds, while Pro-Football-Reference has him at 195 pounds and Clayton lists him at 212 pounds. (There's definitely an issue with figuring out how to handle changing weights of players in our many databases.) So Rice's current BMI probably should land him in the "larger backs" category, but he's listed in the "smaller backs" category. It makes some sense, since that's how Clayton mentions him in the piece. In addition, I also did a version of this study splitting the running backs who by weight rather than BMI (with the line drawn at 218 pounds) and the results were similar. That study would put Maurice-Jones Drew, Frank Gore, Knowshon Moreno, and Rice in the smaller category, with Steven Jackson and Arian Foster in the larger category.

For those curious, here are the numbers for all the backs in the study:

LARGER BACKS vs. 3-4 vs. 4-3
Player Hei Wei BMI DVOA Att Yds Avg BT BT/Att DVOA Att Yds Avg BT BT/Att
33-M.Turner 70 237 34.0 1.3% 87 406 4.67 10 0.11 -1.4% 231 915 3.96 25 0.11
28-J.Stewart 70 235 33.7 8.6% 57 306 5.37 7 0.12 -18.2% 107 434 4.06 15 0.14
32-M.Jones-Drew 67 210 32.9 -0.6% 101 444 4.40 5 0.05 16.5% 180 796 4.42 18 0.10
40-P.Hillis 72 240 32.5 23.1% 126 591 4.69 7 0.06 -5.3% 123 478 3.89 18 0.15
34-R.Mendenhall 70 225 32.3 -0.5% 128 430 3.36 6 0.05 -6.1% 160 692 4.33 21 0.13
27-L.Blount 73 241 31.8 1.9% 56 270 4.82 13 0.23 9.0% 119 655 5.50 20 0.17
21-F.Gore 69 215 31.7 4.4% 49 213 4.35 1 0.02 -13.7% 154 640 4.16 4 0.03
21-L.Tomlinson 70 221 31.7 3.6% 128 513 4.01 4 0.03 -13.6% 73 262 3.59 1 0.01
20-T.Jones 70 220 31.6 -7.2% 66 250 3.79 6 0.09 -30.7% 141 477 3.38 5 0.04
42-B.Green-Ellis 70 219 31.4 9.9% 118 430 3.64 1 0.01 44.4% 76 376 4.95 5 0.07
27-B.Jacobs 76 256 31.2 27.5% 46 292 6.35 7 0.15 14.3% 102 532 5.22 11 0.11
32-C.Benson 71 222 31.0 -10.9% 193 626 3.24 12 0.06 -18.7% 103 367 3.56 12 0.12
23-S.Greene 72 227 30.8 -0.1% 98 364 3.71 5 0.05 18.3% 65 286 4.40 4 0.06
27-K.Moreno 71 217 30.3 14.9% 96 493 5.14 11 0.11 -26.5% 76 244 3.21 8 0.11
AVERAGE       4.0% 96 402 4.17 7 0.07 9.8% 122 511 4.18 12 0.10
SMALLER BACKS vs. 3-4 vs. 4-3
Player Hei Wei BMI DVOA Att Yds Avg BT BT/Att DVOA Att Yds Avg BT BT/Att
27-R.Rice 68 199 30.3 1.8% 155 571 3.68 4 0.03 -0.6% 136 577 4.24 7 0.05
28-F.Jones 70 207 29.7 20.7% 41 201 4.90 4 0.10 4.3% 143 589 4.12 20 0.14
23-A.Foster 73 225 29.7 12.4% 142 648 4.56 16 0.11 22.2% 183 969 5.30 9 0.05
22-M.Forte 73 217 28.6 7.2% 99 425 4.29 3 0.03 -0.6% 123 586 4.76 7 0.06
39-S.Jackson 75 229 28.6 -11.8% 137 489 3.57 14 0.10 -10.3% 149 623 4.18 9 0.06
22-F.Jackson 73 215 28.4 -0.9% 145 549 3.79 12 0.08 2.3% 77 380 4.94 12 0.16
25-J.Charles 71 200 27.9 5.4% 79 426 5.39 8 0.10 42.7% 118 766 6.49 7 0.06
28-A.Peterson 74 217 27.9 15.5% 159 715 4.50 14 0.09 -5.7% 109 473 4.34 10 0.09
20-D.McFadden 73 211 27.8 3.0% 115 584 5.08 9 0.08 6.3% 108 566 5.24 12 0.11
25-L.McCoy 71 198 27.6 24.2% 50 291 5.82 4 0.08 12.1% 139 697 5.01 18 0.13
28-C.Johnson 71 197 27.5 7.1% 121 583 4.82 13 0.11 -15.4% 195 782 4.01 20 0.10
44-A.Bradshaw 71 195 27.2 -11.6% 95 357 3.76 5 0.05 9.9% 180 886 4.92 23 0.13
AVERAGE       4.5% 112 487 4.36 9 0.08 9.3% 138 658 4.76 13 0.09

Posted by: Aaron Schatz on 16 May 2011

47 comments, Last at 18 May 2011, 5:18am by Kibbles

Comments

1
by Sean D (not verified) :: Mon, 05/16/2011 - 4:09pm

Back in the Wade Phillips Chargers DC days, I remember anecdotally saying that the Chargers 3-4 struggled against bigger backs. Jerome Bettis was the main one. But, I know it was a real small sample of backs. There's no way I would've included a Knowshon Moreno-sized back in the big category.

Not that I'm saying anything is wrong with this analysis. Just relaying some thoughts I had back in the mid 2000s.

2
by AnonymousA (not verified) :: Mon, 05/16/2011 - 4:11pm

I would say "larger" probably means "more weight", not "more BMI". A better data presentation might be just four graphs:

1) Weight vs. 3-4 DVOA
2) BMI vs. 3-4 DVOA
3) Weight vs. 4-3 DVOA
4) BMI vs. 4-3 DVOA

Theoretically, if this is a real effect, it would be more extreme at the edges, and show up in a graph; no need to create an arbitrary half-way split.

10
by tuluse :: Mon, 05/16/2011 - 5:14pm

I was thinking the same thing. It's not about how stocky a back is, but just how large. Weight would be a better measure.

25
by Anonymous Coward (not verified) :: Mon, 05/16/2011 - 9:38pm

I would guess there has to be a combination of pure mass and a style that would be factored into how the coach perceives draft prospects.

SJax, Jacbos, Blount, McClain, Hillis , Greene all yes but not an exclusive list

Someone like Foster, Moreno, Tomlinson, or Forte would be a combo player or neither a big back nor a small back

Elusive guys like MJD, CJ, Charles, Rice, McCoy, Woodhead, Washington, Sproles would all be small even though someone like MJD may weigh more than some combo backs

3
by BobTheMan (not verified) :: Mon, 05/16/2011 - 4:14pm

The Totals in the summary chart do not match the totals in the detailed chart at the end of the piece.

6
by Aaron Schatz :: Mon, 05/16/2011 - 4:35pm

Oops! You are right, I cut-and-pasted incorrectly. I'll go fix that.

4
by Independent George :: Mon, 05/16/2011 - 4:32pm

All I know is that in Madden, I liked sending big backs up the middle with a lead blocker vs a 3-4, while sending small backs outside on 2 TE stretches/counters vs the 4-3.

9
by tuluse :: Mon, 05/16/2011 - 5:13pm

That's because Madden does a terrible job simulating 2-gap line play.

29
by Peter Freaking King! (not verified) :: Tue, 05/17/2011 - 12:37am

shit data and the resulting theories based on analysis of shit data are either going to be:
a: Shit
b: Lucky Coincidence

So Madden is more useful in this case.

36
by Independent George :: Tue, 05/17/2011 - 7:17am

That's just crazy talk! Next you're going to be telling me that the dime isn't the best run-stopping defensive set!

5
by Podge (not verified) :: Mon, 05/16/2011 - 4:35pm

A very interesting set of information, although I'm not sure that drawing overall conclusions, but for how different players do against different schemes. Its interesting that 2 similar backs, Chris Johnson and Jamaal Charles are basically completely the opposite in their split against 3-4 or 4-3. Charles is way better against 4-3, while Chris Johnson is way better against 3-4.

And BJGE is basically completely the opposite to Jamaal Charles in running style, but has a similar split in 3-4 to 4-3.

One thing that I noticed - it seems that a lot of backs seem to be better against the same kind of D that their team plays - backs from 3-4 schemes often seem to do better against 3-4, although that's by no means 100%. Or probably even 50%!

I think this is interestingly descriptive, but probably not particularly predictive. I'm curious if there's any correlation from year to year in how a back performs against different Ds.

7
by Aaron Brooks' Good Twin (not verified) :: Mon, 05/16/2011 - 5:01pm

There's also a difference between running against Pittsburgh's 3-4 and Arizona's 3-4.

8
by andrew :: Mon, 05/16/2011 - 5:11pm

I get the impression that John Clayton was grouping AP in with the "larger" backs, while here is grouped with the "smaller" backs.

11
by Intropy :: Mon, 05/16/2011 - 5:21pm

Isn't is a problem that DVOA is already adjusted for defense?

12
by tuluse :: Mon, 05/16/2011 - 5:40pm

I don't see why. If you didn't account for quality of defense, then you would just be measuring that instead of the difference between 3-4 and 4-3.

13
by QQ (not verified) :: Mon, 05/16/2011 - 5:43pm

From a theoretical stand point it makes sense that a Big Back would do better vs a 3-4 (Smaller/Faster front 7) where a larger DL has been removed for a smaller LB and that a Smaller Back would do better vs a 4-3 (Bigger/Slower.

In regards to the study, I would rather see it studied at the team level. For example, did GB struggle more vs Big Backs or Small Backs. Then once all 32 teams have been done to see if any trend emerges. The problem in comparing Peterson's perfromance vs GB and vs DET is that the strengths/abilities of the personnel on both team's is radically different and makes the comaprisons difficult.

16
by JonFrum :: Mon, 05/16/2011 - 6:10pm

Regarding the size difference: at least in Belichick's 3-4, he's using a nose tackle that's heavier than most 4-3 tackles (do you really think Wilfork is 325?) and he prefers his ends to be 300+lbs. Compare R. Seymour to D. Freeney listed at 268. Plus, the linebackers need to be bigger than most 4-3 linebackers in order to take on linemen. It's the difficulty in finding big OLBs who can drop back and cover that's been killing the patriots in recent years. So you replace a lineman with a smaller linebacker, but you make both linemen and linebacker bigger to make up for it.

21
by Xtof (not verified) :: Mon, 05/16/2011 - 8:12pm

I didn't think that 3-4 LBs had to be bigger to take on linemen, I thought they had to be smaller and faster to compensate for the lack of speed on the D line? Either way I figured that both 3-4 and 4-3 front sevens would come to the same weight anyway. Further research required I guess.

22
by dryheat :: Mon, 05/16/2011 - 8:26pm

In a 2-gap 3-4, an OLB is essentially a 4th lineman, albeit one standing up. Belichick likes his OLBs to be 6-4 and 250+ with sub 4.6 speed. As somebody said above, that's a small population.

27
by Aaron Brook's Good Twin (not verified) :: Mon, 05/16/2011 - 10:33pm

That's a TE.

26
by tuluse :: Mon, 05/16/2011 - 9:48pm

I don't think it's as simple as 3-4 vs 4-3.

The Ravens use 4-3 lineups that are huge, the Colts use 4-3 lineups where no one weighs more than 260lbs.

14
by notque_deeproute (not verified) :: Mon, 05/16/2011 - 5:46pm

I run a simulated football game, http://deeproute.com and I love articles like this.

It may not help me on the specific question asked, but seeing the average broken tackle numbers per 4-3 and 3-4 is awesome. I can implement that in our football game today.

Less Broken tackles in a 3-4 than a 4-3, but sort of shocking to me, a 3-4 gives less yards per carry?

Currently we have the opposite implemented. Is this a normal trend? That 3-4 defenses actually allow less yardage running in general? I always thought a 4-3 was better against the run.

23
by dryheat :: Mon, 05/16/2011 - 8:30pm

One isn't better than the other against the run by virtue of being a 3-4 or 4-3 -- In design, they're equal. The chief advantage of the 3-4 is that you get another linebacker on the field, presumably a better athlete than the defensive lineman he's replacing. But some of those Tampa-2 4-3s, like Chicago, have lighter, quicker DTs that compensate.

47
by Kibbles :: Wed, 05/18/2011 - 5:18am

There's a big risk in seeing "the 3-4 defenses in the league today are better at stopping the run" and concluding "3-4 defenses are better at stopping the run".

Thought experiment: imagine every team except for Pittsburgh switched back to the 4-3. What do you think the average performance against 3-4 defenses would look like in that hypothetical world? I would posit that the average performance against 3-4 defenses would be way, way worse than the average performance against 4-3 defenses. Now, imagine Pittsburgh switched to a 4-3, but Buffalo switched back to a 3-4. In that hypothetical, I think it's pretty clear that the average performance against the 3-4 would be way, way better than the average performance against the 4-3. Average performance against a certain type of defense is more a function of which teams are running that defense than it is of any "decided schematic advantage" that said defense might offer.

15
by drobviousso :: Mon, 05/16/2011 - 6:06pm

1) I would not call AP or Steven Jackson "small". A middle size should be excluded here, maybe, just to look at the more extreme examples.

2)And I think this is understood, but we need to normalize for defensive quality. Running on the AFC north's 3 3-4s is probably harder than the NFC west's 2.

17
by DisplacedPackerFan :: Mon, 05/16/2011 - 6:43pm

Glad you did this, now I don't have to. I wondered about Peterson specifically like you said who seems to do just fine vs the Packers 3-4 and since Clayton talks about his numbers declining, which from stuff I've been poking at recently here, I knew wasn't the case. Sure his traditional stats have been falling, but most of the advanced metrics for him are on the upswing. Rushing DYAR has gone 121, 165, 229 the last 3 years; DVOA was -0.3%, 3.2%, 10.6%; though yardage has gone 1760, 1383, 1298 which is why Clayton says he has declined. As a note his receiving has improved greatly the last two years as well. I didn't have a good way to look at the splits though.

Of course I would also consider Peterson a big back not a small back. But I guess that's because he is taller, not heavier. Though I think of anyone above about 215 and taller than 6 foot as big. The sub 6 foot above 215 I end up thinking of as "power" backs I wonder if height actually matters on this splits and not weight or BMI. Or if the real measure of a "big back" is above certain thresholds in height AND weight like my mind tends to do.

So with my odd categories that I had to put numbers too (and then don't always fit the backs in there)


Big (72 or over, over 215)
Player vs... 3-4....4-3
Hillis......23.1....-5.3
Blount...... 1.9.... 9.0
Jacobs......27.5....14.3
Greene......-0.1....18.3
Foster......12.4....22.2
Forte....... 7.2....-0.6
Jackson....-11.8...-10.3
Peterson....15.5....-5.7


Small (less than 210 or sub 70)
Player vs... 3-4....4-3
Rice........ 1.8....-0.6
Jones.......20.7.... 4.3
Charles..... 5.4....42.7
McCoy.......24.2....12.1
Johnson..... 7.1...-15.4
Bradshaw...-11.6.... 9.9


Power (sub 72 over 210)
Player vs... 3-4....4-3
Turner...... 1.3....-1.4
Stewart..... 8.6...-18.2
MJD.........-0.6....16.5
Mendenhall..-0.5....-6.1
Gore........ 4.4...-13.7
LT.......... 3.6...-13.6
Jones.......-7.2...-30.7
Green-Ellis. 9.9....44.4
Benson.....-10.9...-18.7
Moreno......14.9...-26.5


Unconventional
Player vs... 3-4....4-3
McFadden.... 3.0.... 6.3

I know that last category is small but then I don't expect there to be a lot in that range. I would bet Ryan Grant and James Starks of the Packers would be in there too. I just tried to put numbers to the 4 categories I kind of have running backs in.

I do see a trne of "power" backs doing worse against 4-3 with MJD and BJGE being the exceptions. "Big" backs I don't see any trends 4 showed better vs 3-4, 4 against 4-3 (though Jackson was pretty much the same). Though if you take out Jackson they were all pretty much "positive" vs 3-4 and a couple were negative vs 4-3, but I don't see a rough trend. The "small" backs generally showed better vs 3-4 with 4 of the 6 being better. The more "oddly sized" McFadden didn't really show anything.

Dunno, I don't think a two way split works for backs, I really do think a 3 or 4 way split makes more sense. I might not have the boundaries defined right. But the Jim Schwartz quote that started it all may have some legs if what I have as "power" backs are what he considers "big" backs. Clayton fumbled all over it, but that isn't surprising.

18
by Karl Cuba :: Mon, 05/16/2011 - 7:12pm

BMI doesn't seem right for this, I just don't see how Steven Jackson is a small back, he's bloody massive. I'd have said over 230 is a big back.

I also wonder where the stats came from. How can you have watched Frank Gore all year and decide that he only broke 5 tackles? He broke two on the run where he broke his hip and then broke three more on his next three carries, on a broken hip.

Could FO use its statistical tools to award players Madden-style badges: tackle breaker, reciever, short yardage expert, consistent ground gainer etc. Maybe not, or at least work on the names. It would be a quick way to categorise how a player is productive. You could categorise other positions too.

19
by Anonymous2000 (not verified) :: Mon, 05/16/2011 - 7:16pm

1-gap vs. 2-gap doesn't strictly follow with 3-4 vs. 4-3. It would be good to see some accounting of that variable.

20
by andrew :: Mon, 05/16/2011 - 7:32pm

A big part of Peterson's DVOA rating the past few years has been his fumble rate. How do fumbles correlate (if they do) to 4-3 vs 3-4? and small vs large?

What would the numbers sans fumbles look like?

24
by Anonymous Coward (not verified) :: Mon, 05/16/2011 - 9:20pm

LOL at MJD being a large back and SJax being a small back. I'm sure that's exactly what Jim Schwartz thinks when he thinks about the size of those two.

28
by Andrew Milne (not verified) :: Mon, 05/16/2011 - 11:05pm

As a Giants fan, the numbers seem to fit with what I saw and with the hypothesis. Jacobs' average and DVOA go down vs 4-3, Bradshaw's goes down vs. 3-4. But watching, the difference is between an inside runner vs an outside runner (in very broad brushstrokes). Of course, that relates to weight (and speed, which relates to weight...)

30
by Reinhard (not verified) :: Tue, 05/17/2011 - 1:50am

4-3 is based on penetration into the backfield, forcing their players to cut outside, where the fast LBs can always chase them down. So big backs can't get going against them, they have to change direction behind the LOS. Kind of like how NY held Peterson in check that way. More risk-reward; they can get you behind the LOS.

3-4 is based on containment. The three DL have to react oppositely to the offense. If they try to double you down, you have to anchor. If someone is downblocking you, hold your ground and try to push back.
The (bigger) LBs are the same. The OLB have to set the edge, anchoring down against their widest player, like the FB, T, pulling G, or a TE. The ILB are usually bigger although I'm not really sure why.
3-4 chokes away all the daylight.

So if you are "run to daylight" there is more room in the 4-3.

31
by Stewart (not verified) :: Tue, 05/17/2011 - 2:46am

That's a very good qualitative look at it. You're also speaking of true 3-4s, not the hybrid fronts run in places like Arizona which are 1-gap schemes eerily similar to a 4-3 under front with a stand up rusher. The comparison is between philosophical ideals, not the actual pragmatic schemes run across the league. Compare, say, the Patriots 3-4, which isn't the best but is certainly the most traditional representative of the scheme. Very heavy front, weak presence in the gaps but excellent in contain and pursuing to the runner. Less TFLs, but less long gains too. An ideological 4-3 is more like Chicago, using 1-gap responsibilities and I think leading the league this year in TFLs.

34
by tuluse :: Tue, 05/17/2011 - 5:08am

In 2001 the Bears ran a 4-3 featuring the talents of Ted Washington, Keith Traylor, and Philip Daniels. All 3 went to start for 3-4 defenses, and Washington and Traylor played 2-gap nose tackles in 3-4 schemes.

It was very much a 2-gap defense, despite having 4 players lined up in 3 point stances. So I feel just looking at 3-4 vs 4-3 isn't that instructive here. 2-gap vs 1-gap would be much more interesting.

37
by Jimmy :: Tue, 05/17/2011 - 7:41am

Don't forget that the LE was Brian Robinson, last seen playing NT for the Cardinals. What an absurdly big line.

32
by Simeon Rice (not verified) :: Tue, 05/17/2011 - 3:41am

At last someone knows how to coach the 4-3. I go straight into the backfield, miss the runner, and let the OT i was lined up against have free reign on the small 4-3 LB!

Seriously though, 4-3 run-d is about 1 on 1 Gap Control. DL in A & C gaps will try to keep B gap closed by not allowing their opposing linemen to move them laterally. This allows the smaller LBs to get to the ball without big DL hitting them. Its not about getting into the backfield, hence my sarcastic name and opening line. If you're still unaware of your error, a story on Simeon Rice's deficiencies can be found through the link attached to my name.

AND YOU ALWAYS PURSUE OUTSIDE-IN BECAUSE HELP IS COMING FROM THE INSIDE.

never played a 3-4 so i don't know if you're an idiot or trolling.

33
by Load (not verified) :: Tue, 05/17/2011 - 4:36am

Depends on the 4 man line scheme you run- not everything is pursue outside in, some are actually the opposite, and run a spill concept. This was made popular with Jimmy Johnson at the U, where he took slow LBs and put them at DE, and slow DEs to DT.

The spill concept that some 4-3s run have the DEs crush down, and the LBs scrape over the top. The theory behind this was the get the runner moving east/west, where it would be slower to get yards, and also funnel him towards the safeties that would be filling the alley.

A good example of the difference is in the way they play the Power (Gap down by the OL, FB kickout, BSG pull):

Contain 4-3
PSDE when he sees the EMOLOS block down will step up the field and attack the outside shoulder of the FB coming for him, trying to force the run inside.

The PSILB (say Mike in an example, with Sam and the TE playing with one another, say like in an under front) will attack the outside shoulder of the pulling BSG, keeping that outside arm free, forcing the run back inside

The BSILB (say Will) will play the inside shoulder of the pulling lineman, where in theory the RB will have to insert, as Mike has taken away his outside.

The Mike/Will combination is often referred to as "boxing" the RB, as each LB will play a side.

Spill 4-3

PSDE when he sees the EMOLOS block down, will step down (coaches refer to this funnily enough as "block down, step down rule") and use a wrong arm technique (where you attack the kickout block with your outside shoulder, not your inside), forcing the pulling lineman and RB to bounce a gap wider

PSILB will scrape over the top of this (remember, it should be happening tighter than the example above) and attack the inside shoulder of the pulling BSG, forcing the RB to bounce a gap further over

BSILB will fast flow (so flip his hips and run, rather than shuffle and press) when he sees power away, and play over the top of the Mike and DE, making the tackle outside D gap where the RB is running lateral. In theory.

There are strengths and weaknesses to both: in a contain style, if the back breaks a tackle, he is running straight gashing you for yards. The other issue is that the ILBs much take on an OL's block with some degree of force, requiring more power. The advantage of it though is that it'll limit the big plays (in theory) and can handle a back who is too quick for you.

The advantage of the spill method (which is rarely understood by the casual fan) is that the ILBs (in particular the BSILB) can fast flow to the ball, meaning they run over the top of the OL's blocks before they can get to them, meaning they run away from the big guys, and straight to the ball carrier. The disadvantage is that if they take a poor angle, or the RB is just too damn quick, the play goes for miles to the outside.

It really is a misnomer that 4-3 teams are strictly a 1 gap D- sure, you can all be assigned a gap, but what do you do with a stretch type play where the FB is creating a new gap? If anyone is interested (lame?) enough, Rex Ryan spoke at the 2008 C.O.O.L clinic on his D (it's available on DVD from their site) and spoke about if you all just play your one gap, it'll be a long night. Watch a good defensive team playing the run, the linemen will stay square for as long as possible, allowing them to play both their gap, and "fall back" (Ryan's term) into the backside gap of the run fit. Sure, two gapping teams do it all the time and don't look to penetrate (penetration is another way to demand a double team... but that is for another discussion), but don't think that 4-3 teams have one gap, and that's all they care about.

35
by tuluse :: Tue, 05/17/2011 - 5:33am

I wish I could rec you or plus 1 you or something.

I'm pretty sure the Bears run the spill technique, using Charles Tillman's excellent tackling skills on one side and Lance Briggs on the other, and letting Urlacher use his incredible speed and smarts to patrol the middle. This also puts a lot of pressure on the MLB to make the correct read from what I can tell. One way teams have had success running against the Bears is using counters, misdirections and traps.

Just one thing, I think I deciphered all your acronyms, but a few took me a while. EMOLOS = end man on line of scrimmage, right?

38
by Load (not verified) :: Tue, 05/17/2011 - 8:19am

Yeah, EMOLOS is end man on line of scrimmage. Sorry, I forget there isn't a universal coaching language!

You'll find that there is a good chance that the Bears run both systems at different times, for the main reasons:

1) The spill concept is often tied to coverage where there is a safety filling the alley (as you don't want to be spilling the ball to nobody!) so if they are playing some sort of quarters coverage behind it, then they may not spill. I don't watch the Bears games, so I'm not sure whether they still play their cover 2, but if so, you may see the CBs playing that fill role, as they will be bumping then playing the flats, so are in a perfect position to play the run.

You'll find some great examples of the spill concept run by Nick Saban's defences, as they like to play either cover 3 or cover 1 robber, which has one of the safeties rotating down into the alley, so works perfectly. The CBs and the FS play the pass first, and the SS can rotate down and become the 4 LB, and the alley player on his side (they usually rotate down to the wide side of the field, as the SS is a stronger tackler in space, but they don't always because of my following point)

2) In the NFL, if your opponent can predict exactly what happens, they will abuse you. Compared to college where teams don't have as much time to install as many schemes, in the NFL it wouldn't surprise me if they sometimes box the power/iso and sometimes spill it. The reverse of this can be seen in offensive line protection schemes- many colleges only run one or two protections, whereas in the NFL you need multiple protections to ensure you can't be out schemed.

You are correct with the misdirections often causing issues for spill teams, as their LBs are doing the fast flow techniques, which can get them into trouble if they make the incorrect read, or there is counter/split flow.

39
by Big Frank :: Tue, 05/17/2011 - 10:07am

This is not the kind of analysis I want to see on this site. Honestly, I expect a lot more from you guys. This is purely speculative, anecdotal, subjective crap and it means absolutely nothing.

40
by bill (not verified) :: Tue, 05/17/2011 - 10:15am

Sir -

Sometimes the best part of a site like this is the ability of management to prompt interesting discussion. At those times the burden rests with the comunity to educate each other through dialogue, inference, and extrapolation, among others. I'd say we rose to the challenge quite nicely, and are indebted to the author for affording us that opportunity.

Bill

41
by Kal :: Tue, 05/17/2011 - 11:06am

You might try another split: good 3-4s vs good 4-3s. There's probably a lot of noise in the bad teams one way or another that doesn't really add anything; if you play a bad defense it probably doesn't matter what scheme you're using.

But if you only look at the teams with DVOA above 0% on defense, it might be a bit more illustrative.

Still, the problem is with the BMI, I think; that doesn't accurately describe a back's style except in really extreme cases.

42
by BigD1770 :: Tue, 05/17/2011 - 11:35am

Do zone blocking schemes have more or less success vs 4-3 or 3-4 defensive fronts than power-gap blocking schemes? If so, should that also be accounted for?

43
by Aaron Schatz :: Tue, 05/17/2011 - 11:54am

Two quick notes.

1) This was basically just meant to take a quick look at the numbers to see whether Clayton was right that numbers show obvious struggles by smaller backs against 3-4 defenses. Not meant to be a major study or anything like that.

2) That dude in the bodybuilding ad on the left side of the page is seriously freaking me out.

44
by BDAABAT :: Tue, 05/17/2011 - 1:17pm

Would bet that the defensive scheme isn't as large a contributing variable to overall RB success as is offensive line talent/scheme...think Bronco's effective implementation of zone blocking to consistently create 1000+ yard rushers from pretty much anyone in the backfield.

See also changes between Ravens RB performance in 2009 vs. 2010. O-line personnel changed in 2010 with a major drop-off in rushing performance.

Bruce

Acquired sig: Never let your mind remain so open that your brain falls out.

45
by dudes (not verified) :: Tue, 05/17/2011 - 2:43pm

Wait a second... are you trying to tell me that a John Clayton column contained a dubious, half-baked assertion and demonstrable falsehoods? Get right out of town.

46
by CWP (not verified) :: Tue, 05/17/2011 - 9:49pm

Plotting weight vs (3-4 DVOA - 4-3 DVOA) from the table and running a linear regression on the results shows a positively sloped line going about -5% to 15% as weight increases with an R^2 of 0.04 (not a whole lot there). This seems to indicate (however weakly) that larger backs do better against a 3-4 scheme. It dosen't show that the smaller backs do more poorly against a 4-3.

Post new comment

The content of this field is kept private and will not be shown publicly.
  • Web page addresses and e-mail addresses turn into links automatically.
  • Allowed HTML tags: <a> <em> <strong> <cite> <code> <ul> <ol> <li> <dl> <dt> <dd>
  • Lines and paragraphs break automatically.

More information about formatting options