clock menu more-arrow no yes mobile

Filed under:

A closer look at the epic coaching battle that was Pete Carroll vs Bill Belichick

Two of the best coaches around always make fertile ground to study. What did they try? What worked? What adjustments did they make? 

NFL: Seattle Seahawks at New England Patriots
Pete Carroll asks Bill Belichick if he loves Brock Coyle so much why doesn't he just marry him.
Greg M. Cooper-USA TODAY Sports

Not that it matters, but Richard Sherman grabbed Julian Edelman's shoulder.

The far more interesting part, however, is what New England did on that play, and how Seattle responded.

The Patriots lined up RB James White wide right, with Rob Gronkowski and Martellus Bennett in the slot. Edelman lined up in the backfield, then motioned to the slot on the left side. Mismatch creation at it's finest and funnest. If a defense doesn't play its cards right, a severe and vulnerable mismatch is very palpable here.

Especially when New England spent the whole drive in 12 personnel to keep the Seahawks in base. Or, said differently: the Patriots decided they'd prefer Seattle have Brock Coyle on the field instead of Jeremy Lane.

More about Coyle soon.

How does Seattle respond here? Most of your favorite defensive Seahawks were aligned differently than you're used to. Bobby Wagner runs over to cover James White wide. KJ Wright takes strong-side under against Bennett, while Kam Chancellor aligns over the top of both TEs. Richard Sherman aligns like a weak side linebacker on the left side. Edelman's side. This leaves Coyle as the Mike, ready to shoot the gap for a run play, or drop into underneath zone for a pass.

So what happens? Edelman gets the better of Sherman on the shallow out, but Sherman gets to him well before the first down marker, making a remarkable stop. The mechanics of both their bodies strongly suggest an incidental but severe facemask infraction.

Part of me would love to take a moment to explore the epistemology posed here to the officials, and the a posteriori inferences seven referees must make to officiate a game of 22 men, five violent seconds at a time, as though league rules and the physics of reality allow for officiating to be predicated on deductive reasoning. But I'll save that for a rainy day and get straight to the Q.E.D.

Not that it matters. Just a creative set up for a shallow out.

The next play was the one that punctured Gronkowski's lung. I take no joy from such a frighteningly violent development, but Earl Thomas III being one of the smaller humans on the field and Gronkowski being one of the largest makes it astonishing. But mostly I'm astounded by the incredible coverage Chancellor provided on the play.

Anyway. Thomas knocking Gronk out of the game for five plays also put Coyle on the sideline. That's weird because Seattle didn't put Coyle on the field to cover Gronk. So what's up?

Coyle's busy night

Brock Coyle played 34 snaps. Rather than playing to their own strengths, they identified the third LB spot as a relative weakness for Seattle's defense, compared to Jeremy Lane in nickel.

New England did what Buffalo did before them. I don't know how much the 2-back sets and FB lead run game suits Buffalo's offense and personnel compared to 3WR sets, but Bill Belichick was watching:

"This is kind of a unique situation," he explained. "We’re pretty far ahead on Seattle. We’ve had a chance to work on them all week. We’ve done all their games up through the Saints game, so there’s no more to do, and there won’t be any more to do until Monday night. I think Monday night will be a good opportunity to kind of take a look at the game against Buffalo with a little bit of a blank slate.

"Buffalo is a team that we know well from just having played against them (a 41-25 Patriots win) and being a division team, so I think when we see what Seattle is doing, we’ll be able to figure out why they’re doing what they’re doing against Buffalo, what they perceive as an area to attack or what type of game plan to utilize," Belichick added. "I think it will be interesting to watch it in reverse – to watch Buffalo, a team that we know well, attack Seattle. We’re a different team than Buffalo, but still, there are certain things that carry over so we’ll be able to watch how they attack Seattle and kind of do it on a timely basis as opposed to going back and looking at it after the fact."

"We’ll sort of watch it unfold and picture plays or situations, plays and calls that we’ve talked about making in certain situations, and then when those situations come up, see how it would hypothetically match up against Seattle," said Belichick. "This is a little bit of a different chance to do it. As I said, I think that makes it a little bit interesting, so yes, we’ll watch the game."

Not that Coyle was a major factor in Seattle's trouble with LeSean McCoy and Tyrod Taylor, but it worked pretty well for the Bills. The advantage didn't materialize substantially for New England, and late in the third quarter, they switched to 12 & 11 personnel for a few drives which gave us a long spell of Jeremy Lane snaps. By the middle of the fourth, they went back to trying to exploit Coyle and finally got some traction on their final drive. Just a glimpse into the coaching ways of Belichick. Adjustments came before halftime. They tried a couple different strategies, then when it mattered, picked the one they thought worked best.

Getting ahead of Seattle's zone runs

Chris Long got penalized for being offsides, but he wasn't, he just timed the snap. It was really easy. The Seahawks unfold it all blatantly for everyone to see, most snaps.

Mark Glowinski looks back at Russell Wilson. Wilson pumps one foot. Glowinski relays the signal to Justin Britt, who then dips his head down, then up, then snaps. It's very predictable.

It might be necessary for a team that just can't handle how on fleek Wilson's cadence game right now, what with all the false starts and all. Seattle's improved their standing from fifth most false starts last year, to sixth this year. So the lack of hard counts in this game is an OL mitigation that you'll soon learn is part of a bigger trend. But this mitigation doesn't make the OLs job any easier. Because defensive linemen like Long have no trouble timing the snap.

It also allowed New England to try an interesting thing. They didn't have much luck, but other teams might: immediately before the snap, they'd shift their DTs one gap over. They did this five times.

Glowinski's relaying of Wilson's step-signal allows this to work. Alan Branch (now YUGE at 350 pounds!) shifted from 3-tech to tilted 1-tech, Malcolm Brown or Vincent Valentine from 1-tech to 3-tech on the playside. The effect of this is to get bodies in front of the zones to seal off the front side, which should coerce a cutback. It also gives the backside defender some lateral space to evade the cut block, so they can clean up the cutback.

But by coincidence, the Seahawks unleashed their schematic surprise at the same time: pulling linemen. RG Germain Ifedi cuts Branch while Britt and Glow pull to the outside.

Unfortunately for Seattle, Trey Flowers does a great job of blowing up Glowinski's block here and stopping CJ Prosise for no gain. On one of the Patriots' other DL shifts, it came against a pass play, so the impact was moot. On another, they apparently guessed wrong, as the run went to the other side. So it didn't factor much, but on another day, with another team, maybe in our division, it could.

This power pull here is much better executed (Jimmy Graham notwithstanding) so it deserves a bit more attention. Doug's block is superb. Will Tukuafu's down block on Jabaal Sheard is, too. George Fant and Ifedi swallow those massive tackles, Valentine and Branch, exactly as they're supposed to. Garry Gilliam pulls, and the seam to the Mike, Dont'a Hightower, is clean, but he doesn't substantially block him and he doesn't need to. Glowinski's pull is well-executed and Britt's is tremendous.

Pulling is harder for the center, as he has ball-snapping responsibilities. Chip Kelly pulled center Jason Kelce a lot in Philadelphia, and it was pretty impressive. Britt's pull here is on Kelce's level. He also has to block a LB, Elandon Roberts, rather than a corner, Malcolm Butler.

Manufacturing protection and pressure

On the first pulling gif above, the cut block stemming from the guard instead of tackle is interesting. We saw this on the other side, from Glowinski on another play. This is a continuation of schematic mitigations for the weaknesses of the OL. Since preseason, we've seen Seattle use cut blocks by both tackles on passing downs, probably more than two dozen times this year, though not as much recently.

But check out Britt's rollout here. This is not a mitigation. This is a creative wrinkle they have used at least twice before, going back to last year.

They're slide protecting to the right, here, with Jimmy Graham chipping & releasing. Britt & Ifedi double Branch, then Britt peels off and rolls to the chipped edge defender. The Seahawks actually do this again two plays later, because New England's 2-gapping G.O.U.S. (Gentlemen of Unusual Size) left Britt with nothing to do in the middle so often. This is likely a halftime adjustment and not a part of the game plan.

Some examples of actual, designed mitigations: Seattle used Tukuafu or Graham to wham block two times apiece from the H-Back alignment in this game. A wham block is when you intentionally leave a guy on one side of the box unblocked, and someone crosses the pocket from the other side to smack him as he enters the hole.

The Seahawks also used four naked bootlegs to protect Wilson. They probably would have liked to do that more in the past few games if his health would have allowed it more. Naked means Wilson rolled out with no blocker to move laterally with him to protect him. Here's an interesting example, a non-naked example, where Graham stays in to block, and crosses the pocket. He would have whammed if anyone needed whamming.

This is play action to the left, rollout to the right. This helps the pass protection because the defensive linemen track the play action and get into the gaps to each OLs left. As Wilson rolls out to the right, to get around a block, the DL would need to rotate back around to the OL, or cross their face. Generally that's really hard.

On the naked bootlegs, though, when the rollout itself and not any offensive tackle produces some of the best edge protection Seattle's enjoyed all season, this is not a very encouraging thing. But I give our maligned coaches Darrell Bevell and/or Tom Cable credit for fabricating some good pass protection by hook or by crook, as well as the new creativity in run blocking.

The interesting counter to the Seahawks’ established OL problems was how the Pats planned to attack. For the first 20 minutes or so, they stuck to 4-man rushes. By late in the second quarter, they began to drop eight into coverage and only rushed three. They did this six times. On two snaps, they actually started with only a 2-man rush! On the first, a linebacker assigned to Prosise rushed after he stayed in to block; on the second, the two DL rushers were paired with a delayed corner blitz.

New England didn't abandon the 3-man rush, either; they tried it a couple times late in the fourth quarter. Their first 3-man rush resulted in a sack, a coverage sack, by Rob Ninkovich, so that might have emboldened them. I don't know if this was their game plan or if it was an adjustment. I think it's a bit of a reflection of what they thought of Seattle's OL. They thought they could produce sufficient pressure with fewer men. It also reflects respect for Wilson. Not many QBs face a lot of eight-man coverages. Peyton Manning probably faced it more than anyone. You gotta be able to manufacture some pressure, of course, and the Colts' OLs often weren't strong. The Seahawks got a spotty OL paired with a great QB, too, and the best coach in the game thought he needed more men in coverage than rushing, to beat him.

So let's not let the win and the effective offense obscure the fact that Seattle's not out of the woods yet, with their OL. The line's been mitigated significantly. Other reasons for improvement are Wilson's healthier, Prosise's blocking and running helped, and New England's pass rush wasn't great to begin with.

Personnel notes

The OL's been better at pass protection than run blocking. In the past few years it's been the reverse. The biggest pass protection improvement has been more seamless passing off of defenders to one another. On Prosise's 18-yard reception in the first quarter, Chris Long took a 2-gap stunt that got picked up by Ifedi while Britt and Glowinski handled the other two rushers.

George Fant lets his feet set at the bottom of his mirror slide too often. This makes him vulnerable to rips and clubs to get past him even though he's in decent position, due to balance. Trey Flowers got him to get his feet crossed. It wasn't pretty.

Maaan, Jermaine Kearse had a super bad game.

Running back thoughts:

CJ Prosise is a complete running back. I didn't expect the kind of physicality in finishing off runs that he displayed.

The “late Christine Michael”'s chief strength was his incredible burst through the hole. What kept him from being an elite back is lack of patience, mostly. Decision-making isn't always great, either. Sometimes he takes questionable cuts. Way too often, he took the cut that he couldn't make. He takes cuts Barry Sanders couldn't make! So he ends up on the ground without anyone tackling him way too often. He could still improve, but I suspect this will always be a part of his game no matter where he lands.

Thomas Rawls is a better and more impactful runner than Prosise, but I don't expect he'll have a long career. I think he compares more favorably to Marion Barber than Marshawn Lynch. I'm not calling him injury-prone, but just that kind of physicality out of that kind of frame doesn't make one's career long for the NFL. We should appreciate it while it's still here because it could be over in just a couple years.

I have a conditioned response when I see Alex Collins' dreads in the backfield. I expect something bad and disappointing to happen. So far I don't think that's been proven wrong.