A long-winded post comparing Miamit to good teams.

Nice write-up, and I definitely appreciate the time it must have taken to look into all this stuff.

 

Incredible write up. Well done.

Side note: I truly believe we've had horrible YAC receivers (outside of Landry) since Tannehill has been here. I think that greatly limits his YPA.

 

Let's think about what goes into improving the YPA number. Obviously big plays is what pushes that number up... and how do you get big plays?

A) Pushing the ball downfield.
B) Receivers getting more YAC.

Let's delve into each of these: A) How do you push the ball downfield?

Well, you need:
(1) Time for the play to develop so that (2) a QB can throw an accurate deep pass and that (3) the receiver can catch it.

If we delve even deeper into this:

1) How can you buy time for the play to develop?

I) You need an offensive line capable of protecting the QB for the length of time needed. AND/OR
II) You need a QB capable of buying time with his legs.

I think both "I" "II" are pretty shaky right now.

CONCLUSION: In order to fix this component we need to get better OL play and/or Tannehill to develop better pocket presence skills and escapability.


2) How does a QB throw an accurate deep pass?

I) Well first of all, he needs the arm strength to be able to throw the ball.
II) He actually needs to attempt a long pass.
III) He needs to deliver it accurately.

The thing with Tannehill is that his YPA isn't limited by physical abilities like other weak-armed QB's ( e.g. Alex Smith). So "I" isn't an issue here. I think "II" is something that can be worked on. I don't know how our offense's deep throw attempts compare to others but I'm pretty sure it's well below average. You can argue that the line doesn't give Tannehill time to throw the deep ball but I'm pretty sure there are plenty of plays where the option to throw it deep is there, he just chooses not to. And "III" is a work in progress, we will know as the season progresses if he has improved on it.

CONCLUSION: In order to fix this component we need to attempt more deep passes throughout games and Tannehill needs to develop better accuracy on downfield throws.

3) How can the receiver catch the ball on downfield throws?

I) He needs to have enough speed to get open.
II) He needs to be able track and adjust to the ball in the air.
III) He needs to be able to actually catch it.

"I" is something we saw a lot of last season with Mike Wallace. We don't know if we'll have a whole lot of it this season, but we'll see, maybe Kenny Stills (but please, don't want to get into the Wallace debate). "3.II" and "2.III" are inversely proportional to each other, IMO. The more accurate the deep ball, the less 3.II is important and vice-versa.I think we have the receivers that bring "II" and "III" this season, more so than last season.

CONCLUSION: I don't think we'll have a problem this component this season, but it still remains to be seen.

As for B) Receivers getting more YAC, I didn't expect this post to be as long as it was so I'm running out of time to continue this analysis. But I think we have the receivers for this to not be a problem, I'll just leave it at. The only problem is that when A isn't completed very often, the defense creeps down closer to the line and B also becomes harder to achieve.

So, all in all, the things we must work on in order to get better at YPA are:

- More deep pass attempts.
- Better accuracy when attempting the deep passes from the QB.
- Better OL play.
- Better pocket maneuverability from the QB.

Overall, I think most of these things are fixable, though most are harder to fix short term (as in this season). The above order is how I would rank them from easiest to hardest to address. I think the thing that would net us the most immediate results is to actually attempt more downfield throws. Simply put, the more we try, the better chance of making them, and the better we get at completing them as well.

 

I agree with this post except one of the biggest ways to increase YAC is broken tackles. We've had terrible receivers at doing this since RT has been here. If as a team we increased our broken tackles ranking we'd be top 5 offense. This year I expect different things because I feel that all the receivers here, now that Matthews is getting more time we will see an increase in YAC.

 

Thing is, it's pretty hard to break tackles when you're catching the ball 5-6 yards from the LOS. You've got CBs, LBs and Safeties around you, so even if you do manage to break one or two tackles, chances are you aren't going to go very far (we see this all the time from Landry). If you catch a ball 15-20 yards out and you break just one tackle, you have a chance to take it to the house.

 

OP, fantastic!

So he's good enough, right now, to win a SB, using a formula that is tried and true, and not an outlier...that's good news and confirmation of what some already believed.

Let me pull out my troll-prevention spray: "THAT DOES NOT MEAN WE DON'T WANT AND/OR EXPECT IMPROVEMENT!" Let me repeat that before I get called a softener or champion for mediocrity, "THAT DOES NOT MEAN WE DON'T WANT AND/OR EXPECT IMPROVEMENT!"

I think one of the main ingredients for that formula, though, is the coaches of those teams. I don't believe Coyle gets it done, and I don't believe Philbin notices that. I don't know if there's metrics on this, but my feel is that in the post-season, you don't have time to give away a half of the game before you adjust and care about stopping the other team...that goes for Lazor too, at times, but not as much as Coyle...or it could be ALL Philbin...no way to know, it's just how it comes across.

The more the passing game improves, the less of the other stuff needs to happen to get it done. I don't believe we've seen his ceiling, and I don't believe it'll be this year either. He's a slow/steady improving type, obviously.

 

I think there needs to be some research done where the majority of the NFL's and Tannehill's passes are caught and then how much YAC is gotten from those points by other teams and ours.

Also, screen passes say hello.

 

Screen passes are designed and have blockers in front of the targeted receiver (hence the name "screen pass"). Catches made a few yards from the LOS are much harder to turn into bigger plays than ones caught farther out in front.

And I'm speaking relative to last season and not necessarily Sunday's game... I think they will push the ball downfield more often in other games, FWIW. They don't even need to be the 30+ yard bombs that people clamor over. I'm talking 20 something yards, much like the one completed to Jordan Cameron which we all know Ryan's more than capable of. Hoping that with DeVante Parker we will see these types of passes attempted – and completed – more often.

 

These statistics are all well and good but he hasn't won a Superbowl so he sucks

Sent from my SM-G360T1 using Tapatalk

 

No way bruh. He iz da best eva!!1!

 

If Trent Dilfer can win a super bowl RT can get in the hall of fame.

 

7.6 ypa average for the super bowl winning QBs, 6.8 for Tannehill.

That cannot be ignored. Last year Tannehill was in the bottom 5 I think, in distance travelled per pass attempt. Wilson's was even lower, and he had a higher ypa, but lower completion percentage, meaning his receivers and RBs were better at yac. Baldwin averaged about the same yac per catch as Landry (Landry had 20 more catches) so the increase is due to how much they throw to Lynch, with 37 catches, 11 yac per catch.

http://www.sportingcharts.com/nfl/stats/yards-after-the-catch/2014/

Miller had almost the same amount of catches, but his ypc is only 6.5 which kind of sucks for a RB. If you go down the list, most RBs are usually higher. Not blaming Miller, but that's where it's at.

But in any event, Tannehill has never had a high YPA through college and the pros. That has to change. What if it's ball placement affecting the yac? Play design? Offense? We'll find out. He supposedly has better yac receivers now right? So can we see a nice bump to 7.5 or more by the end of the year? When he throws deep, his completion % has fallen down. When we shortened the field, it went up. Both hit the ypa.

It's all about the YPA!

And no, you don't up the YPA to up the yardage, look at Tony Romo who threw 435 passes or something silly, and 3,700 yards. But they balanced it with an effective run game (finally) and look how far they went.

Top 10 passing, effective run game, and decent defense, and strong play in the playoffs can get you to the promised land.

I would also expand the analysis to anyone who gets to the Championship games at least. Sometimes it's just an unlucky bounce that keeps one team from going. But all that get there are usually worthy.

 

The problem is, Philbin hasn't really shown any propensity to favor a balanced attack. Not here at least, and someone said tracing back to GB but that was McCarthy's offense, but perhaps he was heavily influenced by it.

 

Hey Greg, good job trying to find out what makes a SB team!

However..

I gathered the data myself and I think your conclusions about run defense and pass offense being what's most important for a SB team won't hold up in general. Why? Because as far as I can tell (unless I put in the data wrong), your results are being skewed by the team that lost the Superbowl, as well as by the "top 10" threshold.

Try this yourself with the data: do the same analysis but JUST on the team that won the SB. You'll find that there are basically the same number of top 10 pass offenses (measured by total yards) as top 10 run offenses among SB winning teams in the years you looked at. On the other hand, the SB losing team has a really skewed stat: far more top 10 pass offenses than top 10 run offenses in those years.

Similarly, though there are more top 10 run defenses among SB winning teams in those years than top 10 pass defenses, the difference is relatively small. For SB losing teams the difference is huge! So your stats are being skewed by the SB losing team.

The other problem is the relatively arbitrary "top 10" threshold. Try looking at how many top 5 rush defenses vs. top 5 pass defenses among SB winning teams there are. You'll find more top 5 pass defenses than top 5 run defenses (the opposite of your theory). Or look at "above average", as in top 16, and you'll find they're equal.

Anyway, like I said I admire your attempt to find some statistical regularity here, but I don't think we can trust your conclusion about run defense + pass offense being the most important because it doesn't hold for SB winning teams.

 

You know, I was getting ready to write "Great Post OP! That was an excellent analysis that was both logical and unbiased." But then you had to go and say what you just said, which values the statistics more than the actual plays on the field. And that kills your entire theory since stats and execution are definitely not the same thing.

How? Let's go back to game 1. Tannehill put up respectable numbers which many here have argued were over-inflated. The national media has echoed that point and Tannehill himself said today-

http://sports.yahoo.com/news/dolphins-ndamukong-suh-calls-play-poor-season-opener-180048933--nfl.html

Now, let's say that he hits that pass to Sims. Or the sideline pass to Stills at the 3 yard line. Or he doesn't fumble just outside the red zone. Pick any one of the three and change the outcome. Statistically, it doesn't do a thing for his overall QB rating except for bringing up a fraction of a point (slightly higher if he got the TD pass to Sims or if Stills ran for a TD after the catch). But it completely changes the game overall and our perception on his performance. More importantly though, it also changes the team's belief in him to deliver when it really matters.

So saying he's "good enough" because he statistically had a good game is complete crap. Mistakes cost us at least 17 points, and the dropped pick-6 could have accounted for another 7 point swing. That's a 24 point differential left on the field for a team that lost by an average of 8 points last year.....and that's the difference of going 8-8 or 16-0 in 2014. So don't make the mistake of reading solely into the numbers and ignoring what your eyes see on the field, because "statistically" Washington won game 1. They still get to wear that big "L" regardless though, because the stats do indeed lie.

Great post, by the way, you just over-did it with that final comment.

 

Hard to take serious when you make stuff up to try (and fail) to support your point.

Speaking of "complete crap"...a fraction of a point you say?

Had he hit the pass to Sims, their rating would have been 108 instead of 93. TDs are HEAVILY weighted in passer rating computations, despite some around here thinking they don't. Fairly big difference stat-wise, ALSO a big difference in the complexion of the game, his play, and the myriad conversations about it.

 

Sample size is a real concern. I found a site that shows us offensive passing vs. rushing efficiency for all playoffs teams from 2003-2012:
https://thepowerrank.com/wp-content/uploads/2014/01/nfl_pass_rush.png

That shows that if you take into account all playoff teams, passing matters and rushing doesn't, relatively speaking. So I'll agree with you on that now.

OK..

The defensive side of the ball is trickier (and probably more work-intensive). I found a graph showing all the playoff teams from 2006 and their rush defense ranking here:
http://archive.advancedfootballanalytics.com/2007/03/importance-of-run-defense.html

It shows basically no correlation. But I can't find one for a lengthy period of time (and same goes for pass defense), so we may have to do the work ourselves, unless someone here can find such a graph (and keep in mind we need rush/pass defense ranking or efficiency measured by yards or yards per carry vs. making the playoffs, NOT playoffs wins or something else..). I don't have much time right now, but I'll start gathering those stats later this evening (Greg you should too so we can compare and get to the bottom of this with adequate sample size).

Also, there are two other problems with the argument that rush defense matters for winning that we need to resolve:

1) Having a statistically good rush defense may to some degree not be a "cause" of winning, but an artifact. Specifically, teams that are ahead (esp. late in the game) tend to have to defend more against the pass, not the run, so they should have inflated rush defense rankings. That does NOT mean they are as good against the run as their ranking would suggest, meaning you don't need to invest as much in it. Thus, ideally the stats only include 1st-3rd quarter rush defense stats. I won't go that far right now (harder to collect) but it's important to keep in mind.

2) It almost defies logic that if passing is what's important, that you should build a strong defense against what's not important: the run. That doesn't make sense logically, so something's fishy here.

Anyway, we should first get the stats on rush and pass defensive rankings for all playoff teams over a decent period.

 

There's nothing to disagree with in your analysis...I think it's awesome and very insightful. So don't think I am saying anything different here. The only point I was trying to make is this-

A quarterback goes 25/31 for 402 yards, 3 TDs and 0 INT, which is an almost perfect quarterback rating (the troll here will tell us the exact number so he can feel like he's contributing). However, those numbers tell you nothing about the actual score of the game....that QB's team could have lost 27-48 to another QB that went 22/42 for 388 yards 2 TDs and 2 picks. The numbers themselves will never fully tell you the complete story with a quarterback since there are other factors within the offense involved (like the great run offense you pointed out).

So I don't think there's a baseline for what's "good enough" in a quarterback, and I definitely don't think there's one for "bad enough" either. You look at QB's like Mark Rypien or Eli Manning and they had one or two amazing seasons with a career of mediocrity. Heck, look at names like Flacco, Kaepernick, Rex Grossman, Rich Gannon or Jake Delhomme- none of them were truly special for entire seasons. But they put their teams in good positions down the stretch and closed out ballgames when it really mattered.

Is Tannehill more skilled than several of the pros I just named? I think he is. But saying that 4k yards or a certain number of TD's guarantees that he'll be ready once the playoffs arrive is just insane. Because there's no direct correlation.

Heck, look at Griese back in the magical 70's...he sat the bench for a stretch of the perfect year and almost didn't start in the playoffs. Statistically, neither of the Fin's QBs met your average for the season yet there they were...the best team of all-time. And that's all I meant by don't let the stats lie; they just don't tell the entire story.

 

Amen.

 

I don't think we can say that with certainty at this point. If a running back averaged let's say 3.2 yards per carry (another measure of efficiency) rather than the 4.5 or so you'd hope for from someone you were counting on to help carry the team, might it say he's a bad running back? Possibly, sure. By the same token, Tannehill's (and/or the team's) inability to have a 7+ YPA during his tenure could possibly mean he's an inadequate QB. And without a 7+ YPA, the odds are heavily against its being a playoff-caliber (10+ win) team, unless it achieves a rare level of greatness in all other phases (run and pass defense, and running game).

 

And that again would be a function of offensive YPA, i.e., the ability to put up points offensively and make other teams one-dimensional as they try to mount comebacks via the pass.

 

OK, I compiled the stats:

Average rush defensive ranking of all playoff teams from 2003-2014: 12.1
Average pass defensive ranking of all playoff teams from 2003-2014: 15.1

So there is a difference it looks like and Greg's right about rush defense being more important than pass, though I don't know how big a difference 3 levels in rank is.

I'm still confused by this though.. if passing is more important than running (there is a much bigger difference there), then why invest in run defense more than pass defense??

 

Ya at this point it could go either way. I personally feel the arrow's pointing up, but there's no proof of that from a long-term consistency standpoint.

But week 1 is always a bad game for RT, so I don't put much stock in their YPA after 1 week. There were several QBs this past week whose teams had sub-par YPA (below 7.5 IMO), but are considered at minimum franchise QBs (for comparison, RT (6.65 won), Luck (5.0 lost), Flucco (3.66 lost), Peyton (4.38 won), Wilson (6.12 lost), Brees (7.4 lost), CKap (6.35 won), Alex Smith (7.36 won), Scam Newton (5.65 won), Eli (5.36 lost), Bradford (6.46 lost)).

So basically a 50/50 proposition last week...5 won and 6 lost.

 

Because once you shut down the run, the offense becomes a lot more one-dimensional and the D can adjust to make plays (taking away the slant routes, jumping passes in the flats, etc.). But it also means that you have linemen hitting those gaps hard...which means that they're pressuring the QB more often as well.

 

You could just as well make the offense one dimensional by shutting down the passing game, taking the QB (and the passing game) out of it even more than by shutting down the run. So that won't explain it.

The only theories I can come up with that make sense are:

1) The run defense rankings truly are more an artifact of better teams being ahead in the 4th quarter, and thus playoff teams aren't as good at run defense as the rankings suggest.

2) These are rankings for regular season only! That means that maybe it makes sense to have a strong run defense to get INTO the playoffs (because we already know that most teams that don't get into the playoffs have a weaker passing game), but not necessarily to succeed IN the playoffs.


Both theories can be tested. Theory 1 can be tested by comparing run defense rankings in the 1st-3rd quarter to overall run defense rankings. Theory 2 can be tested by comparing the pass defense rankings of winning playoff teams vs. losing playoff teams, for every playoff game. Both require a lot of work to gather the data. Not sure I'll go after it, but if I do I'll probably try testing theory 2 since that would be more informative. Note however I've already tested theory 2 in one key situation: SB winning teams do have higher average pass defense rankings for those 12 years than SB losing teams, so maybe that's it?

 

My comment definitely does explain it. To shut down the run game, you control the line of scrimmage. Poof, it's shut down. Detroit was tops against rushers last year, allowing only 1,109 yards and 8 TD's.

To shut down the passing game, your entire linebacking corp and secondary need to deflect or intercept every pass thrown throughout the entire game. That's impossible against five moving targets- the best pass defense last year, for example, gave up 2,970 total yards, 61.7% completion ratio and 17 TD's through the air.

So yes....a great rushing defense can make an opposing team one-dimensional. A great passing defense still gives up 185 yards a game, 6.3 yards per attempt and over 1 TD. A great rushing D gives up an average of 69 yards and zero TD's....huge difference.

But then again, you're talking 4-6 players making up the core of that rush D, with 6-8 for that top pass D (with the linebackers overlapping in both). So controlling the line and shutting down the run is by far the more sensible goal in the NFL.

 

Interesting.. hadn't considered the resources needed to shut down one or the other. Of course resources matter but resources alone won't answer the question because it doesn't tell us how the relative value of passing vs. running, or the relative decrease in opponent passing/running efficiency, enter into the equation.


You know what? I just tried something out with a simple/crude mathematical model that might answer the question (in favor of targeting run defense btw..).

Let's suppose the following:
R = rush yards
P = pass yards
v = statistical value of passing relative to running (we know from the above graphs that v > 1 because passing is more valuable than running)
h = multiplicative fractional decrease in efficiency of opponent passing or running (h must be a number between 0 and 1).
Y = total yards gained

OK..

Assume for simplicity that rush yards gained and pass yards gained are totally independent of another (not technically true, but it's a simple approximation). Then we can represent the statistical value of total yards gained when the defense does not preferentially target opponent passing or rushing as follows:

Y = vP+R

If the opponent preferentially targets passing, then we have:

Y = hvP + R

and if they preferentially target rushing, it's:

Y = vP + hR

Now, a yard passing = a yard rushing (independent of anything else), so the "P" and the "R" are really just place holders.. we can replace them with anything, including a number, such as "1" (we are only interested in the effect of the weights h and v). So you have:

Y = hv+1 for when the opponent targets passing, and
Y = v+h when the opponent targets rushing

Remember here that v>1 and h is between 0 and 1. To see under what conditions targeting passing is better than targeting rushing, we need to solve for v in:

hv+1>v+h

The solution is when v<1, which by assumption is impossible (we assumed that statistically speaking passing is more important than running, or v>1). If we ask when targeting rushing is better, you have to solve:

hv+1<v+h

and that does have a consistent solution: v>1.

What that means is that according to a simple mathematical model where you assume rushing and passing yards are obtained independent of one another and there's statistically speaking greater value in passing, it's always better to decrease the efficiency of opponent rushing.

Still very counter-intuitive..


EDIT: Man.. am I stupid!! The equations are right but the interpretation is wrong (wasn't looking at it from the opponent's view). The only consistent equation is hv+1<v+h, meaning if you target rushing (v+h) the OPPONENT will have more yards than if you target passing (hv+1). LOL.. so the conclusion IS intuitive: if resources don't matter, targeting passing is better than targeting rushing!

 

That was really long. Damn dude do you have a life? Maybe you should publish a book about it.

 

I have a life. Right now I have some free time because I more or less finished what I set out to do today. Anyway, the problem is still vexing.. why target what your opponent isn't that good at?

I'll admit KeyFin came up with another possible reason (resources), but it still doesn't make sense.

 

I don't have time right now to verify your formulas, but look at it from an entirely different, on-the-field perspective.

Defenses, by design, play the pass every single down and they hope the front 3-4 can stop the run. Because think about this- even when the opponent has 1st and goal from the 1 yard line, defenses have over half the team spread out to defend the pass and take away the middle of the end zone. So clearly, the focus is always on passing.

Here's what should happen in the average game; those linebackers start on 1st and 10 reading the quarterback's eyes, following the ball and deciding within a split second whether to stuff the gap, cut off the edge or drop back into coverage (in the flats or wherever). This is BY FAR the hardest position in football because it's one of the few positions where they could be 10+ assignments for every single play, and it's also the position that has the most direct impact. If the linebacker makes the correct read, gets in position and makes a play, then it's likely a positive down for his team.

And here's what really happens; the linebackers get a "feel for" the offense and they start to jump plays, coming on blitzes (by design or by instinct), dropping into coverage or moving to one side of the field. When you have an awesome defensive line, then you don't have to worry about cheating up to the LOS or reading the QB as precisely, because you know that those front gaps are likely covered. So you jump routes more often and take a lot more chances...all because you control the line of scrimmage.

When the opposite is true though and a team is running free like we saw in Washington, then the linebacker has to stay at home and always play run-first. And that's just the linebackers; what about those free safeties and corners when a team is running at will? They have to read the quarterback just as much while moving at full speed in initial coverage...which is why we occasionally see wide open receivers off of a play action. When you're controlling the line of scrimmage though, then these headaches do not exist. Given that passes are much more dangerous to a D than a run play, you want to minimize your run exposure with as few bodies as possible.

So again, a great run defense can dramatically boost a passing defense, while the opposite is almost never true. That's why you play the pass every down and find specialists to help stop the run up front. It's a critical component of any top ranked D.

 

Well this I disagree with.

You're right about the effect of a great DL on freeing up LB's for pass protection (same argument I was making with the hoped-for effect of Suh), but it's also true that great CB's free up defenders for run defense. That's exactly what someone like Revis at his best can do. So I don't agree with the bolded part.

Also, there are many situations where the defense primarily plays the run, certainly often seen on 1st and goal at the 1.

 

If you want to determine whether run or pass defense is more important, take a look at the players teams are paying the highest amounts of money on defenses. Those are pass-rushers and the corners. And that stands to reason, since passing efficiently on offense is what wins in the NFL. If you can decrease the opposing team's pass efficiency with pass defenders (DEs and CBs), you stand a much better chance of winning.

 

I think in general you're right (the most valued positions all revolve around the QB, either helping him or hurting him with the passing game). Here's a graph with all the average salaries + range:
http://www.businessinsider.com/nfl-highest-paid-positions-2014-9

But the riddle still remains as to why the average pass defense ranking is lower than run defense ranking for playoff teams. I would expect otherwise.. wouldn't you?

 

What are those rankings based on? The most appropriate numbers to consider would be opposing teams' YPA (for pass defense) and opposing teams' YPC (for run defense). That would remove the confound of numbers of run and pass attempts. Also, league rankings may be a poor choice, as well, since they may be skewed. A better figure would be a team's degree of deviation from the league norm (either up or down, if at all).

 

Totally agree with all that (esp. teams' deviation from the norm, but that takes more effort so I didn't do that), so if the reason for the rank difference is found in the rank methodology, please point it out.

I used total yards rushing/passing (either over a season or per game). Reason for that is that I got too many teams with "equal" rank when you use YPA/YPC etc.. There were some years where like 4 teams would be ranked at #6 and then the question becomes whether the difference in rank was due to using #6 or #9, etc...