Upon Further Review- NC State

View as article

Lance Roffers

Junior
Joined
Feb 23, 2018
Messages
1,989
Welcome to an abbreviated version of Upon Further Review, when D’Eriq King made time stand still and transformed a legion of Canes fans back to the glory days with a performance that has been talked about all across the country.

Run fits have been an issue all season. Here, you see Bradley Jennings has already vacated his gap by getting outside of the RB and giving him a cutback lane. He needs to flow inside-out to defend this run.
1.png


I’m not sure what keys Jennings is reading here. He completely sprints out of his gap on this play again leaving an easy cutback lane for the RB. This is truly poor LB play on these two plays.
2.png


Blades tries to jam with left hand as he should here but isn’t physical enough. When he turns his shoulders towards the sideline here and allows the receiver to get back inside of him he’s already toast on this deep ball.
3.png


More LB issues on this play. The QB faked a jet sweep and just keeps the ball. Jennings actually didn’t fall for the fake at first but didn’t realize the QB had the ball and just ran out of the way here. Gurvan Hall is completely fooled as well. If I were an OC against this defense it would be eye candy and misdirection all game. They just cannot identify the ball and read their keys. Frierson makes an outstanding play or this is a huge play.
4.png


Any defense that asks McCloud to turn his back and run to find a TE in the seam is probably going to end up poorly. McCloud was a tick late off the snap, then a tick late to get his head around and the ball goes through his hands and the TE makes a nice catch. It was honestly nice coverage and a great play if he just finishes this. Bolden then uses his shoulder rather than wrapping up and misses the tackle.
5.png

6.png


Roche makes a nice play to knock this pass in the air and it goes right through the hands of Jaelen Phillips. This is what is called turnover variance, and is essentially the unpredictability of turnovers. Miami could’ve had interceptions on back-to-back plays and didn’t convert either of them. What matters is putting yourself in a position to have the opportunities because the variance will swing your way at times and you’ll make these catches. NC State appears to have been more fortunate than great early in this game.
7.png


The second TD for NC State is a nice play from their WR, who didn’t show hands too early and fought through contact and some poor technique from Ivey, who stopped his feet on contact and never tracked for the ball after getting in trail position. (Not pictured)

Offensively, Miami has been to reach this backside block on a slip screen exactly zero times this season. This backside block is crucial to the success of the play and for whatever reason, Miami hasn’t been able to get to it.
8.png


Put this on teaching tape for how to press a hole before you bounce it outside. This run actually goes outside of RT, but you’d never know it here. Harris does an excellent job of pressing this hole and getting defenders to engage.
9.png


As I wrote, this one is abbreviated from some of my marathon reviews, so I’m going to fast forward to when the game is 31-24 and Miami is trying to hang tough.

Miami gets a 1st down on a square-in thrown perfectly by King. Next play is another example of how pressing the hole that your OL is blocking to sets up cutbacks. If Harris runs to the cutback immediately, the defenders stop flowing and right there for the cutback gap. A jump cut behind this block from Clark opens a hole to run through. For me, the biggest transition for HS RB’s to college is to understand how important pressing the initial hole is to the health of the overall running game.
10.png


Harris was pretty good in this game at finding lanes and picking up more yards than what were blocked initially. Here, Harris presses that stretch play but then finds the cutback lane to pick up a huge 1st down.
11.png


Couple of things here, Knighton saved this play by making a nice blitz pickup after initially stepping up. Next, King stands in there knowing he is going to get blasted by the defender getting past Jarrid Williams inside. I saw that on film while at Houston routinely, but it doesn’t hurt us here as King completes a deep shot to Pope down the sideline.
12.png


We’ve been waiting for them to fake that bubble to the boundary and throw the go and they do it at the right time here. King sees a single-high safety so he knows that S will have trouble getting over to defend this play. King sets it up with a full shoulder and hip fake to the bubble and Pope sells it well by running the bubble the same way he always does. CB steps up on the pump and it is over at that point.
13.png


Kickoff return for a TD and Parrot is slightly out of his lane and that’s all it took. He needs to be a step to his left and then he can defend the cutback and the returner bouncing it. He gets too close to the other defender and misses the tackle and it’s a house call. Ragone missed a tough tackle, but it’s on Parrot, not Ragone.
14.png


Fumbled exchange. If King holds onto this, he is out the gate up top.
15.png


Hall lets the WR cross his face to easily on this play, as he should be sitting on the left side of this route with a left-handed QB. Jennings falls for the fake, when he needs to get depth on this play to help with crossers. Look at the giant hole Jennings leaves in the MOF here. That’s just too easy for the QB to complete passes.
16.png


Send Jennings on a blitz and this is too easy for the offense. RB releases into the flat on 3rd & 4 and you ask your WLB to get all the way across and defend this with your MLB blitzing. There is just no way you are stopping this play with this defense.
17.png


Catch the ball Blades. Bubba, look for the football and not trying to kill the receiver.
18.png


Gaynor definitely heard about this in film study. He got tall, exposed his chest, the NT dogwalked him back into the RB. Harris makes an excellent jump cut and gets outside here. Gaynor has to be better.
19.png


Details make plays work. Coaches say it all the time to players and here you can see how details make this play work. Mallory is going to release into the MOF and they’re going to hit that pop pass again. Here you can see Mallory is setting up like he does every time when he’s run blocking on this play. TE’s or outside blockers will use that inside arm to create contact with the outside blocker. This lets him know you’re there as you’re supposed to be, and it create a radius of blocking for the TE. As soon as he does this, you can see every box defender screaming towards the run. Mallory lets the defender go by him and then releases. If he releases early, he gives away the play.
20.png


I’ve seen this more than once this season and this is what separates Mallory a bit from elite guys at the position. Mallory has to fight back towards the ball, rather than fading on these plays. He’s not getting the calls that he should because he’s fading, rather than causing the defender to run through him. This ball should’ve been picked off by #6. This is still pass interference, even though they didn’t call it.
21.png


Slip screen again goes nowhere, but what I really want to see is Lashlee come back to this play in a big spot. If King pumps this and throws the go down the sideline it is a walk-in TD. The second defender that is deep in the slot jumps the screen and vacates the outside receiver (who normally blocks on this play). That receiver was jumping up and down as he knew they had the play and King didn’t throw it.
22.png


What a man’s play by Nesta. Takes the blocker and the RB down.
23.png


Send Couch off nickel, Brooks jumps the hot, jam the receivers instead of playing Cover-2 shell. This is more of the defense I want to see from Miami. Physical at point of attack. Forcing difficult throws with pressure and utilizing players in their best areas.
24.png


Overview:

Miami goes on to win it as King makes some sensational throws down the stretch. The defensive approach in the 4th quarter was different than we saw for the other three quarters. Miami went away from their base cover-2 shell defense that was getting shredded in the flats and over the MOF. They made Hockman look like a credible QB and that is due to the fact they made things too easy for him. Once they started getting physical with outside receivers, using Brooks in coverage in the MOF (rather than Jennings, who is tentative in the MOF), they started shutting things down. Couch made some excellent plays as a blitzer and it’s a nice play to go to against a team like NC State who has a limited QB.

King made plays that we haven’t seen from a QB in at least a decade, but what really gives me hope after reviewing the film is seeing that Miami really was unfortunate in many respects. There were five balls in this game that were tipped into the air, right in our defenders’ hands, or went right through their hands that weren’t taken advantage of. The sixth was completed and was the game winner. Add to it a kick return TD, when normally Miami hasn’t allowed those this season, and it was fluky for NC State to be in it as they were. That is called variance and Miami was on the extreme end of it in this game and that will regress towards the mean over other games. If that variance was average in this game, Miami wins by 10+ and doesn’t have to sweat.

Love that in spite of adversity, they still found a way to win the game.
 
Advertisement
I understand this notion of turnover variance and TO's effectively being unpredictable, but it does not hold if athletic ability is not equal among the two teams, therefore rendering regression to the mean unlikely if not impossible.

TO variance has to assume equal talent (ceteris paribus), otherwise there will always be bias.

To make an extreme but simple example, if you put 2001 Miami against some lousy high school team, there will never be regression to the mean for turnovers given the talent gap, and therefore TO variance fails to hold water.

Basically, the notion of TO variance is only applicable in the NFL (over a relatively long time period) or equally-matched college teams. Even at that, given the law of large numbers relative to randomness, regression to the mean can often take a while as TO's can be "streaky."

EDIT: For a really good understanding on large numbers and randomness, I highly suggest reading A Random Walk Down Wall Street, which is effectively about finding false trends in large data sets because the law of large numbers provides an environment for unlikely occurrences.
 
I understand this notion of turnover variance and TO's effectively being unpredictable, but it does not hold if athletic ability is not equal among the two teams, therefore rendering regression to the mean unlikely if not impossible.

TO variance has to assume equal talent (ceteris paribus), otherwise there will always be bias.

To make an extreme but simple example, if you put 2001 Miami against some lousy high school team, there will never be regression to the mean for turnovers given the talent gap, and therefore TO variance fails to hold water.

Basically, the notion of TO variance is only applicable in the NFL (over a relatively long time period) or equally-matched college teams. Even at that, given the law of large numbers relative to randomness, regression to the mean can often take a while as TO's can be "streaky."

EDIT: For a really good understanding on large numbers and randomness, I highly suggest reading A Random Walk Down Wall Street, which is effectively about finding false trends in large data sets because the law of large numbers provides an environment for unlikely occurrences.
You realize I’m literally certified in Data Science, right?
 
Advertisement
Great write-up.

Baker is so in over his head at Miami, it's scary.

Manny needs to bring in someone who can inject some true X's and O's into this defense or it will be his undoing.

The "downfall" of 2020; King is saving Baker's *** with monster games like he had last week—and in Louisville.

Miami would probably be sitting at 3-4 right now if Baker's defense didn't have this offense bailing them out weekly.
 
I understand this notion of turnover variance and TO's effectively being unpredictable, but it does not hold if athletic ability is not equal among the two teams, therefore rendering regression to the mean unlikely if not impossible.

TO variance has to assume equal talent (ceteris paribus), otherwise there will always be bias.

To make an extreme but simple example, if you put 2001 Miami against some lousy high school team, there will never be regression to the mean for turnovers given the talent gap, and therefore TO variance fails to hold water.

Basically, the notion of TO variance is only applicable in the NFL (over a relatively long time period) or equally-matched college teams. Even at that, given the law of large numbers relative to randomness, regression to the mean can often take a while as TO's can be "streaky."

EDIT: For a really good understanding on large numbers and randomness, I highly suggest reading A Random Walk Down Wall Street, which is effectively about finding false trends in large data sets because the law of large numbers provides an environment for unlikely occurrences.
Now add a confounder like poor coaching, which, like talent, I suppose needs to remain relatively equal as well. Miami's ****-poor coaching negates any the talent gap in a divisional game like this and thus balances the equation.
 
Now add a confounder like poor coaching, which, like talent, I suppose needs to remain relatively equal as well. Miami's ****-poor coaching negates any the talent gap in a divisional game like this and thus balances the equation.
No. I was talking about the variance of how a ball bounces. Nothing to do with talent bias, it is simply a certain percentage of passes tipped into the air would be expected to fall into a place that it is intercepted. Miami got 1 of 6. How a ball bounces is random and that is what will regress to the mean.
 
Advertisement
imo shaq and pinckney were always really good at attacking the blocker's outside shoulder and not letting plays get outside of them. it was noticeable in these breakdowns over the last few seasons. what i feel like i'm seeing from jennings in this game is a robotic version of that style of play -- attacking the outside shoulder of his blocker even as it ends up w/ him overrunning the correct gap by several yards and letting the RB cut back behind him for a huge gain.
 
No. I was talking about the variance of how a ball bounces. Nothing to do with talent bias, it is simply a certain percentage of passes tipped into the air would be expected to fall into a place that it is intercepted. Miami got 1 of 6. How a ball bounces is random and that is what will regress to the mean.
Okay, but I was responding to @CashMoneyCane who said, "TO variance has to assume equal talent (ceteris paribus), otherwise there will always be bias."

My premise is if you add all potential confounders he's introducing they would ultimately nullify each other.
 
Advertisement
You realize I’m literally certified in Data Science, right?
Nope. But if that's the case I'm assuming you understand exactly what I explained, and am surprised you would make a blanket statement like "That is called variance and Miami was on the extreme end of it in this game and that will regress towards the mean over other games."

If turnovers are truly random (which literally cannot be quantified and therefore not stated with certainty -- read: I don't buy they are), then over the long term, the expectation is that every team will end up with +/- 0 turnovers, again, CP. That's the regression to the mean I'd expect, from instances where you have significant variance from 0 in turnovers.

It's a very interesting statement considering turnovers in this game didn't look remotely like a statistical anomaly (one total, which is probably not one standard deviation over all observable college football for over 100 years), which is typically where you'd expect regression to the mean. For instance, if we actually got all those turnovers you discuss, and were +4 or whatever, then I'd say that's variance, and then I'd expect regression. Or, more specifically, randomness was observed over a small sample size, and given variance, you will have regression over a greater sample size.

Is your implication that our mean turnovers lost(gained) is greater than zero(one), and over what population of games will there be an adequate data set to determine what said mean is? One season, to me, is not enough data, and greater than one season, to me, means all the inputs are not the same and therefore the resulting data is not statistically significant, if not wholly invalidated. And even IF you could look at one season, you'll inevitably have marked issues with talent gap, injuries and all the myriad other factors that are inherent to football.
 
No. I was talking about the variance of how a ball bounces. Nothing to do with talent bias, it is simply a certain percentage of passes tipped into the air would be expected to fall into a place that it is intercepted. Miami got 1 of 6. How a ball bounces is random and that is what will regress to the mean.

OK, yes, randomness of ball bounces are one thing. But in a game where you only actually have one turnover, what regression are you discussing? We'll actually secure more TO's on tipped balls? I feel like there's a lot of further qualifying to do there to have any semblance of significance.

Okay, but I was responding to @CashMoneyCane who said, "TO variance has to assume equal talent (ceteris paribus), otherwise there will always be bias."

My premise is if you add all potential confounders he's introducing they would ultimately nullify each other.

Correct, that's exactly what I'm getting at. You cannot say "turnovers are random" given all the other inputs. You can say "opportunities for turnovers occur randomly," which is as close to accurate as you're going to get, again, given all the other factors.
 
Advertisement
Your writeup gives me hope, Lance. But while I agree Miami gets more of those TOs 9 out of 10 games, and the game isn't as close as a result, I also feel this game exposed defensive scheme and talent weaknesses that shouldn't be there 5 years into Manny's defensive leadership. We made it too easy for NCS to score against us. It will be interesting to see if Manny takes an even more active role for Va Tech.
 
Excellent work. I miss the long reviews though.
I don't mind short and to the point reviews! We all saw poor linebacker play. Missed blocks, Holding and take downs by NC state offensive lineman.I want to see us play a complete game on Saturday. I hope we fire Baker at some point. He has no business coaching on a P5 level. He is awful and lacks the passion to do his job successfully.
 
Advertisement
Back
Top