A Metrics Breakdown of Patterson, Allen, Hopkins and Bailey | FinHeaven - Miami Dolphins Forums

A Metrics Breakdown of Patterson, Allen, Hopkins and Bailey

NUGap

Active Roster
Joined
Mar 7, 2012
Messages
236
Reaction score
14
Finally a breakdown that's more relevant to Dolphins fans than the QB one I just posted. I've stuck this on my site at (http://secondroundstats.com/2013/02/04/tier1-wrs/), but I'm going to post a Dolphins-centric version here.

Similar to the quarterbacks, I went through these player’s games and marked down a variety of factors. I noted where they caught the ball, how many yards they picked up after the catch and more. Statistics don’t tell the whole story, but now when someone tells you that a certain wide receiver is a beast at picking up yards after the catch – you’ll know better. For the purposes of this post, I’ve considered the top 4 wide receivers as Keenan Allen, DeAndre Hopkins, Cordarrelle Patterson and Stedman Bailey (you’ll see why).

Where Are They Catching the Ball?

I promise this will be the only chart like the one in the QB post, the rest aren’t so similar. This represents what zones they caught the ball in, before yards after the catch. Unfortunately, I don’t have the exact routes or what side of the field they caught it on. That will have to wait until the next iteration of this.




  • Keenan Allen lives on the short passes. 63.3% of Allen’s passes were caught within 5 yards of the line of scrimmage. Only 3.33% of his passes were past 20 yards. A low for all WRs I looked at. This may not be a bad thing if his yards after the catch are good. I don't know if that's the playmaker that we want in Miami.
  • Stedman Bailey’s game consisted of a lot of screens. Unlike Keenan Allen though, we see a much more distributed catching range. 18% of his passes were caught in the 11-20 yard range with 10.5% deep catches. A great distribution for the Dolphins offense, someone who can catch the ball at all levels consistently and still make plays.
  • Surprisingly, Hopkins was a major deep threat. This surprised me because I thought of Hopkins as a guy who ran a lot of curls and mid-range outside routes. We see that 70% of his catches were past 6 yards. The highest in the major WRs for this class, excepting Terrance Williams.
  • Patterson’s numbers are just interesting. We don’t see many passes caught past 20 yards, but 33% of his passes were caught in the 11-20 yard range. It’s like he decided to ignore catching the ball in the screen and 20+ yard game and just catch intermediate passes.

What’s Happening After the Catch?

This chart represents the yards from the LOS that they caught the ball before YAC and then the yardage after the catch in the second bar.



  • DeAndre Hopkins shows us how much deeper he caught the ball than the others. On average he caught the ball 12 yards from the LOS, before YAC.
  • Stedman Bailey’s yards after the catch is great. In this class, most WRs YAC hovers around 5.3-5.5 yards. Bailey’s is the highest at 6.24. Even though 33% of his catches were screens, having good yard after the catch skills makes it worthwhile. This seems like he'd be a good fit for Miami's offense, a reason why I like him so much.
  • Allen’s numbers are not so superb. On average he caught the ball 4.57 yards from the LOS. This is 3 yards lower than the next wide receiver (Quinton Patton). Plus his yards after the catch is simply average. This scares me a bit for Miami's offense. We hear about the importance of a playmaker and yards after the catch, yet Allen wasn't making those plays in college? I understand the whole thing about Zach Maynard, but still I'd expect more.

How Did Their Systems Help/Hurt Them?



This one is going to require a little explaining. I didn’t just chart their catches, I charted every pass thrown to each wide receiver. In that, I was able to derive how often a QB targets his number one wide receiver and how often QBs miss their wide receiver. Thus I averaged out the percentage of targets, miss percentage, and average amount of throws per game, to give each WR the same amount of targets. Then I adjusted to see how their season numbers would have been, had they been in an average system

  • Patterson is helped by this the most, by far. First off, Tyler Bray was just bad in terms of missing his wide receivers. However, Patterson was also targeted far less than a normal number 1. Had he been targeted at the same rate, he would have gained 549 yards, for a season total of 1327 yards. There were certainly be no questions about his production with those numbers.
  • Bailey incidentally is hurt by this. This of course is due to the high powered passing offense of West Virginia. I don’t think this drops Bailey’s value at all, because he’d still have 1378 yards, but it shows you the influence of WV’s offense.
  • Hopkins numbers would be obscene with more targets. He’s looking at north of 1600 yards with more targets, this of course is due to his high average catch distance.
  • Allen's numbers remain mostly unchanged. This means in terms of targets and misses, Allen represents the average number 1 WR. I understand Maynard wasn't the best QB and it doesn't take into account sacks and whatnot, but it's a start.

*One final note on all of this. I realize this is imperfect. Would Hopkins have been as much of a deep threat without Tajh Boyd? Thus if he had a different QB, would he have caught as many deep passes with more targets? Possibly. I’m trying to work on a way to solve this, but this is my first pass at the WRs.

I didn't want to make it too long (people stop reading after 1000 words and multiple charts), so there is more data/ charts on the site, like red zone yardage, yardage by down, yardage by quarter and drop percentage. You can find that here:
http://secondroundstats.com/2013/02/04/tier1-wrs/

EDIT: I'll be posting a Tier 2 one later, I've already got everything done for Wheaton, Williams, Patton and Hunter and that'll be up in a few days. I've been getting a lot of requests on some other places, so I'll probably also be doing Cobi Hamilton, Aaron Dobson, Robert Woods, and maybe one more.
 
Last edited by a moderator:
All very interesting. The thing is with Hopkins, I'm not sure his lack of screens was coincidental. I don't find him all that compelling after the catch.

Yet you have a similar issue with Cordarrelle Patterson. Lack of screen passes in the distribution. It would normally suggest Tennessee didn't trust Cordarrelle much with the ball in his hands after the catch. But even a cursory glance at his tape would show that theory to be silly because he was an absolute nightmare to defend with the ball in his hands.

Just curious, do you have all this on spreadsheets somewhere? What did Cordarrelle Patterson gain on average per attempt on screen passes? I ask because in particular when you're looking to normalize everything with that final graph, the production that you would actually add into Cordarrelle's game would probably be the screen production, since that's really what he's missing.

Unless, is that what you did?
 
All very interesting. The thing is with Hopkins, I'm not sure his lack of screens was coincidental. I don't find him all that compelling after the catch.

Yet you have a similar issue with Cordarrelle Patterson. Lack of screen passes in the distribution. It would normally suggest Tennessee didn't trust Cordarrelle much with the ball in his hands after the catch. But even a cursory glance at his tape would show that theory to be silly because he was an absolute nightmare to defend with the ball in his hands.

Just curious, do you have all this on spreadsheets somewhere? What did Cordarrelle Patterson gain on average per attempt on screen passes? I ask because in particular when you're looking to normalize everything with that final graph, the production that you would actually add into Cordarrelle's game would probably be the screen production, since that's really what he's missing.

Unless, is that what you did?

I'm looking at around 12 YPC (including YAC) on screens. For what it's worth, Justin Hunter ran screens around 10% of the time as well (and we're looking at 18% on average for the WRs I did). I'm not sure if it's the Tennessee offense or what? I aimed to get a 6-7 game sampling, or around 50% of their total catches, so the sampling should be accurate - but it could be off.

I didn't normalize it, that was my initial thought. While not an overwhelming endorsement of myself, I wrote a program to collect data faster (graphic inputs, etc rather than just inputting directly into Excel), and it was really rough on the first iteration. Thus, with the WRs (my test group), the data is extremely difficult to sort through and it was almost impossible to pull the data to normalize. The data with the QBs was much better, which is why I was able to post it quicker than the WRs. So the short answer, no it's not normalized, but that would have been the way I would have gone.
 
What I'm suggesting for instance is, take the following.

Let's say you had 8 screen attempts thrown at Patterson. The QB missed 4 of the throws, he caught 4 of them and gained 60 yards. You had 30 non-screen attempts within 1 to 5 yards. The QB missed 10 of them, he caught 20 of them for 200 yards. You had 25 non-screen attempts within 6 to 10 yards. The QB missed 10 of them, he caught 15 of them for 200 yards. You had 33 attempts within 11 to 20 yards. The QB missed 20 of them, he caught 13 of them for 260 yards. Finally you had 4 attempts at 20+ yards, the QB missed 3 of them and Cordarrelle caught 1 for 60 yards.

Let's say you calculate what you call "average" QB performance, and you come to believe that the average QB misses only 1 out of 8 screen throws, only 6 out of 30 throws at 1 to 5 yards, only 7 out of 25 at 6 to 10 yards, only 16 out of 33 at 11 to 20 yards, and only 2 out of 4 at 20+ yards.

Let's say also that you calculate the "average" offensive pass distribution and realize that out of 100 pass attempts, notmally about 18 are screens, 25 are 1 to 5 yards, 23 are 6 to 10 yards, 21 are 11 to 20 yards, and 13 are 20+ yards.

If you were to adjust Cordarrelle's production to an average passer in an average offense, you would be adjusting his yardage total from 780 yards to 1263 yards, without increasing his workload.

See what I mean?

---------- Post added at 04:38 PM ---------- Previous post was at 04:36 PM ----------

I'm looking at around 12 YPC (including YAC) on screens. For what it's worth, Justin Hunter ran screens around 10% of the time as well (and we're looking at 18% on average for the WRs I did). I'm not sure if it's the Tennessee offense or what? I aimed to get a 6-7 game sampling, or around 50% of their total catches, so the sampling should be accurate - but it could be off.

I didn't normalize it, that was my initial thought. While not an overwhelming endorsement of myself, I wrote a program to collect data faster (graphic inputs, etc rather than just inputting directly into Excel), and it was really rough on the first iteration. Thus, with the WRs (my test group), the data is extremely difficult to sort through and it was almost impossible to pull the data to normalize. The data with the QBs was much better, which is why I was able to post it quicker than the WRs. So the short answer, no it's not normalized, but that would have been the way I would have gone.

What do you mean by graphic inputs?
 
What I'm suggesting for instance is, take the following.

Let's say you had 8 screen attempts thrown at Patterson. The QB missed 4 of the throws, he caught 4 of them and gained 60 yards. You had 30 non-screen attempts within 1 to 5 yards. The QB missed 10 of them, he caught 20 of them for 200 yards. You had 25 non-screen attempts within 6 to 10 yards. The QB missed 10 of them, he caught 15 of them for 200 yards. You had 33 attempts within 11 to 20 yards. The QB missed 20 of them, he caught 13 of them for 260 yards. Finally you had 4 attempts at 20+ yards, the QB missed 3 of them and Cordarrelle caught 1 for 60 yards.

Let's say you calculate what you call "average" QB performance, and you come to believe that the average QB misses only 1 out of 8 screen throws, only 6 out of 30 throws at 1 to 5 yards, only 7 out of 25 at 6 to 10 yards, only 16 out of 33 at 11 to 20 yards, and only 2 out of 4 at 20+ yards.

Let's say also that you calculate the "average" offensive pass distribution and realize that out of 100 pass attempts, notmally about 18 are screens, 25 are 1 to 5 yards, 23 are 6 to 10 yards, 21 are 11 to 20 yards, and 13 are 20+ yards.

If you were to adjust Cordarrelle's production to an average passer in an average offense, you would be adjusting his yardage total from 780 yards to 1263 yards, without increasing his workload.

See what I mean?

---------- Post added at 04:38 PM ---------- Previous post was at 04:36 PM ----------



What do you mean by graphic inputs?

Yeah I know what you're saying, and it makes sense. Now that I think about it, I could actually use the data from the QBs I did last year/ this year to do it. I was only thinking about it in terms of using my data from the QBs of the WRs. I may go through and work on that, it wouldn't be too hard since my QB data is pretty easy to use.

That sounded fancier than I meant it to sound. I just meant instead of typing my data directly into excel boxes (like I did slowly last year), I wrote a basic program using button inputs, text boxes, clicking on the yard line rather than entering it with numbers. Just made it faster. But the way it wrote to excel was pretty bad initially, so the data is hard to parse.
 
Oh ok. Well, if you're going to come up with some sort of normalization that involves hypothesizing how all receivers would've produced had they gotten the same distribution of passes and the same number of attempts their direction with the same quality of QB throwing to them (in terms of "misses") then I think it would be a worthwhile venture. You seem to have already done a mass amount of the legwork. The rest seems just formula stuff.
 
Fascinating. I think it confirms my feelings on Allen and why I'm not pushing for us to pick him . In his defence, YAC from a screen or short pass is almost certainly harder won than YAC in space down field.

It also helps support the theory that Bailey and Hopkins would be the two "instant" contributors, as opposed to say Patterson. Really interesting stuff and so much more could be inferred from it with a little more work from your serious raw data.

Great work.
 
Fascinating. I think it confirms my feelings on Allen and why I'm not pushing for us to pick him . In his defence, YAC from a screen or short pass is almost certainly harder won than YAC in space down field.

It also helps support the theory that Bailey and Hopkins would be the two "instant" contributors, as opposed to say Patterson. Really interesting stuff and so much more could be inferred from it with a little more work from your serious raw data.

Great work.

Curious to know what your rationale is for that.
 
Great work man! very interesting stuff. I know its still early in the draft process but I just dont see us taking such a raw WR such as Patterson at #12. I think we go with the best pass rusher available or best DB available. In some drafts WR's are flying off the shelf being selected too early, and other years a WR wont be selected till the 2nd round. I think this years first round is going to be a lot of WR's taken in the mid-late 1st round. All four of the guys used in the metric break down could go in the 1st round, even Terrance Williams and Justin Hunter. And if teams start picking WR's then Miami needs to trade up back into the first and grab one. I would like to draft one of these guys later in the first or 2nd round
 
Great stats and thanks for the time you have put in...I would love to see Allen and Hopkins get drafted by Miami this year.
 
Great breakdown! I really like Bailey and Hopkins...would go nuts if we could get both.

Sent from my DROID RAZR using Tapatalk 2
 
great stuff. would still like to see what allen could do with a competent quarterback and another WR to prevent defenses from doubling and in stanford's case, triple covering him. When you look at bailey, he has austin across from him. when you look at patterson, he has hunter across from him. when you look across at the other WR from allen, he has a guy named chris harper. maynard is also one of the worst Qb's i've ever seen in my life, and that's not an exaggeration.

and to the OP, the last chart. allen actually had 737 yards this year. so with avg qb, he would have 245 more yards?
 
great stuff. would still like to see what allen could do with a competent quarterback and another WR to prevent defenses from doubling and in stanford's case, triple covering him. When you look at bailey, he has austin across from him. when you look at patterson, he has hunter across from him. when you look across at the other WR from allen, he has a guy named chris harper. maynard is also one of the worst Qb's i've ever seen in my life, and that's not an exaggeration.

and to the OP, the last chart. allen actually had 737 yards this year. so with avg qb, he would have 245 more yards?

Forgot to mention that, I just projected Allen out based on his season numbers as if he had played 12 games so he could be compared with the others.
 
Back
Top Bottom