Outliers

[Dantés]

Using metrics like athletic testing can be helpful or harmful depending on how they're used.

Becoming too enamored with workouts at the expense of whether a guy can actually play football will lead to some poor draft picks.

But ignoring athletic testing and drafting really poor athletes will cause just as many problems.

So testing metrics are best applied by ruling out (or least highlighting) the outliers-- the guys who profile physically/athletically in a way that rarely succeeds in the NFL.

If you rule out outliers, the pool of prospects left to you will just carry a higher probability of hitting. That's not to say that an outlier can't hit, but that your odds of success in taking one are poor.

So who are the outliers in this class who teams need to be careful of? Who profiles in a such a way that almost never succeeds at the NFL level?

 

[Dantés]

Devonta Smith didn't participate in Alabama's pro day, but he did reveal that he weighs 170 lbs.

There have been about 270 receivers drafted going back to 1987 who weighed under 180 lbs.

To my knowledge (going back that far, there could be others that I don't know of), five of those players went on to produce 1,000 yards seasons:

Desean Jackson-- ran a 4.40 at 5'10" 169#
Flipper Anderson-- ran a 4.45 at 6'0" 169#
Isaac Bruce-- ran a 4.55 at 6'0" 173#
Keenan McCardell-- ran a 4.56 at 6'0" 175#
John Brown-- ran a 4.34 at 5'10" 179#

Five players in a sample of 270 is a hit rate of 1.9%. And the only two recent cases, Brown and Jackson, were elite speed players (which Smith reportedly is not-- presumably why he's opting not to test).

So a team drafting Smith has to hope that he's the outlier that will buck the trend. I think I would let someone else take that bet.

 

[Dantés]

Another high profile wide receiver, Rondale Moore came in at 5'7" flat, 180 lbs.

He tested just as explosive as he looked on tape-- blazing a 4.29 forty and jumping 42.5".

However, players this short basically never become regular fixtures in an offense.

In my research, I see about 125 wide receivers at the combine since 1987 who measured in below 5'9".

The only ones who became regular players at receiver in NFL offenses to my knowledge:

Tyreek Hill: 68.13" 185#
Jamison Crowder: 68.38" 185#
Marquise Goodwin: 68.88" 183#

That's three guys in 125, 2.4% of the sample, and all of them were 1+" taller than Moore.

5'7" players have generally been gadget/ST players in the NFL. That role has value, and I would draft it, but not on days 1 or 2.

So I would let someone else take Rondale top 100 to try and make him a 70+% snap player at receiver.

 

[GleefulGary]

Are you just turning our conversations into threads now?

Lol.

 

[PackFan2]

Loving this thread Dantes. I've been thinking about this...
One of the biggest factors to the other opposite spectrum of "outliers" are the prospects from small schools. It's crazy but it's true, the NFL Hall of Fame is littered with small school guys (of course a handful from big names colleges ND USC ETC) BUT STILL. I enjoy looking for guys like Shannon Sharpe or John Randle. I take them as case studies.


Now to the other spectrum...
This isn't quantitative but more qualitative per se-- I'm not a big fan of QBs that had a lot of weapons in college (with some exceptions).
Like Mac Jones and Justin Fields.

I'm also iffy on Devonta Smith. He might end up being a one-trick pony or the next Marvin Harrison. Not sure how that skinny body is going to last in the NFL...

 

[ARPackFan]

I think any draft prospect that falls outside the physical statistical norms for their position deserves a a closer look. There are reasons for certain body types to be better suited to a specific position so they can hopefully stand up to the physical demands. I doubt a 250# O-guard would make it in the NFL even if he kicked a** at the collage level. Then again sometimes outliers have long successful careers but as you point out it is unlikely - think of Muggsy Bogues (5'-3") in the NBA for 14 years. I think taking an outlier in the draft is a risk best taken fourth round lower.

This isn't quantitative but more qualitative per se-- I'm not a big fan of QBs that had a lot of weapons in college (with some exceptions). Like Mac Jones and Justin Fields.

Chris Simms (PFT & Podcasts) takes that into account along with throws under pressure when evaluating QBs. This is why he has Zack Wilson (BYU) over Trevor Lawrence. BYU is not exactly a small school but Simms' point is he does more with less and makes more good throws when pressured.

Found this statistics project where a collage senior looked at NFL height & weight averages by position. I don't know what year the student prepared it for but I would imagine it is still relevant. Pages 12 & 13 of the document has the box plots with whiskers (outliers) plotted.

https://webpages.uidaho.edu/~renaes/251/HON/Student PPTs/Avg NFL ht wt.pdf

 

[GleefulGary]

I think not taking an athlete who was incredibly productive against good talent, while having extremely high testing results, all because he’s an inch or two too short, is well, short sighted.

When it comes to Rondale Moore, him being 5’7 or 5’9 doesn’t change his game at all. He’s a slot only, manufactured touch player. He was never going to play on the outside. He doesn’t have long arms, so he will have to find ways to get open bc contested catches are going to be more difficult for him. Well guess what?! He’s good at getting open because he’s a freak athlete! He’s like a supercharged Wes Welker or Edelman! Moore’s issues have more to do with health than his size. So sure, he might fit a database better if he was 5’9, but it wouldn’t change his game in the least bit. As far as evaluating him as a player, I don’t think it matters that much. Now if Moore was a player that in college won on the outside a lot, and used his speed/strength to get open, then I think it’s a different topic. But he didn’t. He was a slot and manufactured touch player. That’s his game. It will continue to be his game. His height (or lack thereof), doesn’t change it. I’d take him round 2, easily. If he was there in the 3rd, I’d sprint it in.

When it comes to Devonta Smith, I mean, what more do you want the guy to do? He beat everybody he played, and it wasn’t close. And don’t give me the “he’s surrounded by talent” BS. That matters for Mac Jones, not Smith. Smith went Mano y Mano with some of the top CB’s in the country the past two years, against players that will excel in the NFL, but people don’t think he can do it in the league? Cmon! He’s a stud. I’ll be flabbergasted if he isn’t good. He’s better than Calvin Ridley, at a minimum.

 

[GleefulGary]

I think any draft prospect that falls outside the physical statistical norms for their position deserves a a closer look. There are reasons for certain body types to be better suited to a specific position so they can hopefully stand up to the physical demands. I doubt a 250# O-guard would make it in the NFL even if he kicked a** at the collage level. Then again sometimes outliers have long successful careers but as you point out it is unlikely - think of Muggsy Bogues (5'-3") in the NBA for 14 years. I think taking an outlier in the draft is a risk best taken fourth round lower.



Chris Simms (PFT & Podcasts) takes that into account along with throws under pressure when evaluating QBs. This is why he has Zack Wilson (BYU) over Trevor Lawrence. BYU is not exactly a small school but Simms' point is he does more with less and makes more good throws when pressured.

Found this statistics project where a collage senior looked at NFL height & weight averages by position. I don't know what year the student prepared it for but I would imagine it is still relevant. Pages 12 & 13 of the document has the box plots with whiskers (outliers) plotted.

https://webpages.uidaho.edu/~renaes/251/HON/Student PPTs/Avg NFL ht wt.pdf

I’ll give Chris Simms credit because he has very interesting takes, and he obviously watches a lot of the tape. He’s always out there a bit on some prospects, but he’s not a Skip Bayless hot take guy.

With that said, he’s wrong about Wilson over Lawrence. He’s just wrong.

 

[Dantés]

I think not taking an athlete who was incredibly productive against good talent, while having extremely high testing results, all because he’s an inch or two too short, is well, short sighted.

When it comes to Rondale Moore, him being 5’7 or 5’9 doesn’t change his game at all. He’s a slot only, manufactured touch player. He was never going to play on the outside. He doesn’t have long arms, so he will have to find ways to get open bc contested catches are going to be more difficult for him. Well guess what?! He’s good at getting open because he’s a freak athlete! He’s like a supercharged Wes Welker or Edelman! Moore’s issues have more to do with health than his size. So sure, he might fit a database better if he was 5’9, but it wouldn’t change his game in the least bit. As far as evaluating him as a player, I don’t think it matters that much. Now if Moore was a player that in college won on the outside a lot, and used his speed/strength to get open, then I think it’s a different topic. But he didn’t. He was a slot and manufactured touch player. That’s his game. It will continue to be his game. His height (or lack thereof), doesn’t change it. I’d take him round 2, easily. If he was there in the 3rd, I’d sprint it in.

When it comes to Devonta Smith, I mean, what more do you want the guy to do? He beat everybody he played, and it wasn’t close. And don’t give me the “he’s surrounded by talent” BS. That matters for Mac Jones, not Smith. Smith went Mano y Mano with some of the top CB’s in the country the past two years, against players that will excel in the NFL, but people don’t think he can do it in the league? Cmon! He’s a stud. I’ll be flabbergasted if he isn’t good. He’s better than Calvin Ridley, at a minimum.

Right, so if Rondale is a manufactured touch player rather than just a fixture who produces within the flow of your offense, you probably shouldn’t take him in day 1 or 2. He has a role, but that role is unlikely to be a 700+ snap receiver.

We can say 2” doesn’t matter, and maybe they won’t, but the plain fact is that it has mattered. If he became a full time receiver at 5’7” he would be the only player to do so in three plus decades so far as I’m aware. Why take those odds high in the draft when you have other good options?

As for Devonta Smith, what I want him to do is to not be in a sub set of receiver prospects, physically, that almost never work out.

The whole point of this is not to say that these guys can’t play or that they’re guaranteed to fail. It’s simply a way of eliminating the highest risk prospects to improve your odds of hitting. Because yeah, Smith looks awesome. And it looks right, on tape, to say that he would at least be as good as Ridley. But Ridley himself was undersized and still twenty pounds heavier than Smith.

You let some other team have the chance of Smith being a stud to avoid the misses, which will always be more prevalent when you’re dealing with outliers. That’s the theory.

 

[GleefulGary]

I don’t like passing on good players. That’s my strategy.

My point on Moore is that when he was 5’8 or 5’9 people had no issue taking him in those rounds. Now that he’s 5’7, it’s a problem? His game hasn’t changed! He’s the same freaking player! He was always going to be a slot and manufactured touch player.

Kevin King, by the numbers, should’ve been a great CB. Met every threshold. But when he steps on a football field, he looks like a chicken with its head cut off. Draft good players.

 

[Dantés]

I don’t like passing on good players. That’s my strategy.

My point on Moore is that when he was 5’8 or 5’9 people had no issue taking him in those rounds. Now that he’s 5’7, it’s a problem? His game hasn’t changed! He’s the same freaking player! He was always going to be a slot and manufactured touch player.

Kevin King, by the numbers, should’ve been a great CB. Met every threshold. But when he steps on a football field, he looks like a chicken with its head cut off. Draft good players.

The point is not that great testing numbers make one a great player. It's that if you use them to rule out guys who fall below certain thresholds, the pool you are left with is a much higher probability pool of players.

Now within that pool that remains, tape study is still paramount. You don't just draft a guy because he's a good tester. You do "draft good players"-- after helping yourself out by winnowing the field a bit.

No one has to agree with me, but my opinion is this: at 5'9", there would be great precedent for believing that Moore could be a full time slot receiver who can do whatever you would normally ask your slot player to do (i.e. play 70+% of your snaps). At 5'7", that precedent does not exist. Hence I would be careful about how high I draft him.

 

[PackFan2]

I think any draft prospect that falls outside the physical statistical norms for their position deserves a a closer look. There are reasons for certain body types to be better suited to a specific position so they can hopefully stand up to the physical demands. I doubt a 250# O-guard would make it in the NFL even if he kicked a** at the collage level. Then again sometimes outliers have long successful careers but as you point out it is unlikely - think of Muggsy Bogues (5'-3") in the NBA for 14 years. I think taking an outlier in the draft is a risk best taken fourth round lower.



Chris Simms (PFT & Podcasts) takes that into account along with throws under pressure when evaluating QBs. This is why he has Zack Wilson (BYU) over Trevor Lawrence. BYU is not exactly a small school but Simms' point is he does more with less and makes more good throws when pressured.

Found this statistics project where a collage senior looked at NFL height & weight averages by position. I don't know what year the student prepared it for but I would imagine it is still relevant. Pages 12 & 13 of the document has the box plots with whiskers (outliers) plotted.

https://webpages.uidaho.edu/~renaes/251/HON/Student PPTs/Avg NFL ht wt.pdf

Loving this. A simple stats 101. As a Statistics Professor, I would say they could have included standard deviations so we can see the spread and range per each position.

I respect Chris Simms opinions -- he takes a lot of heat because he doesn't follow the norm. his take on Kellen Mond made me do a double-take... IMO Mond over Mac Jones.

My WR and OL prospecting have been a bunch of hits and major misses. Not a fan of the size of Devonta Smith.. and like I said I hope I'm wrong. like my poor take on Shenault & Jefferson but its only a year thus far. Claypool and Gab Davis were WR I was pounding the table for.

 

[GleefulGary]

The point is not that great testing numbers make one a great player. It's that if you use them to rule out guys who fall below certain thresholds, the pool you are left with is a much higher probability pool of players.

Now within that pool that remains, tape study is still paramount. You don't just draft a guy because he's a good tester. You do "draft good players"-- after helping yourself out by winnowing the field a bit.

No one has to agree with me, but my opinion is this: at 5'9", there would be great precedent for believing that Moore could be a full time slot receiver who can do whatever you would normally ask your slot player to do (i.e. play 70+% of your snaps). At 5'7", that precedent does not exist. Hence I would be careful about how high I draft him.

I think the reason the precedent doesn’t exist is because there are hardly any prior examples of it.

Rondale Moore is a very unique player, with a unique size. Typically the smaller players are pure track stars that dabble in football, like Trindon Holliday.

I understand why you, and a lot of other people smarter than me, wouldn’t take him high. I just don’t like how people who were high on him when he was 5’9 now won’t take him bc he’s 5’7. I think it’s similar to how people now think Oweh is a 1st round pick bc he ran a 4.4 40. Production still matters at some point.

 

[ARPackFan]

Loving this. A simple stats 101. As a Statistics Professor, I would say they could have included standard deviations so we can see the spread and range per each position.

I respect Chris Simms opinions -- he takes a lot of heat because he doesn't follow the norm. his take on Kellen Mond made me do a double-take... IMO Mond over Mac Jones.

My WR and OL prospecting have been a bunch of hits and major misses. Not a fan of the size of Devonta Smith.. and like I said I hope I'm wrong. like my poor take on Shenault & Jefferson but its only a year thus far. Claypool and Gab Davis were WR I was pounding the table for.

I have to admit my infatuation with statistical analysis. I'm an electrical engineer by education but spent the last 10+ years of my career doing PRA (Probabilistic Risk Assessment) in the nuclear power industry.

The point I would like to make concerning the use of statistics is that it only gives insights into the probability of an outcome and should only be one of multiple tools used to evaluate a player. Devonta Smith may be the next Larry Fitzgerald or Jerry Rice but but based upon physical parameters it becomes unlikely.

 

[Dantés]

I think the reason the precedent doesn’t exist is because there are hardly any prior examples of it.

Rondale Moore is a very unique player, with a unique size. Typically the smaller players are pure track stars that dabble in football, like Trindon Holliday.

I understand why you, and a lot of other people smarter than me, wouldn’t take him high. I just don’t like how people who were high on him when he was 5’9 now won’t take him bc he’s 5’7. I think it’s similar to how people now think Oweh is a 1st round pick bc he ran a 4.4 40. Production still matters at some point.

I think where we are talking past each other is in regards to the bolded statement.

The idea of identifying outliers is not that it makes it clear who to draft, or that it means production/tape doesn't matter.

It's taking the entire pool of draft eligible players (especially those in consideration in the top 100) and identifying the ~5% of those players who profile in ways that rarely, if ever, hit in the NFL in the way that you'd want that player to hit if you took him in the top 100.

So you're still left with ~95% of the class, and those guys should not be ranked by mere athletic testing, but by their entire profile, with tape being the most important part of that profile.

Think of it this way:

Players who look very, very good on tape in college football routinely fail in the NFL.

Why does that happen?

Well it's more than just these reasons, but sometimes it's injury, sometimes it's off-field, sometimes it's athleticism or strength or size issues.

Teams regularly eliminate or significantly drop prospects because of durability concerns, trying to avoid guys who will bust because they can't stay healthy.

Teams regularly eliminate or significantly drop prospects because of off-field concerns, trying to avoid guys who will bust because they're knuckleheads, or just no committed.

And teams do (and should) eliminate or significantly drop prospects because they fall below thresholds of athleticism, size, and/or strength that tend to translate to the NFL. When I talk about "thresholds" in this sense, I don't mean only selecting for elite athletes. I mean a low bar, below which you almost never see NFL success.

So for example, Asante Samuel Jr. is a 5'10" 180# average athlete. I probably wouldn't draft that at #29, but I would definitely not scratch his name. Because nothing about being 5'10", 180 lbs, or an average athlete has said you are significantly unlikely to succeed in the NFL.

Similarly, Trevon Moehrig came in at 6'0" 205# and tested like a good, not great athlete. I would 100% consider him at #29 because his tape is reportedly so good.

All of this to say, the point is not to blindly take the best athletes, but to be very cautious about the ones that so rarely translate.

 

[GleefulGary]

I understand that.

I think with Rondale Moore, for example, it’s important to note the distinction that the reason players his size haven’t succeeded is because the sample size is very low, and not because a great number of them have failed. There’s a difference there, at least to me.

 

[Dantés]

I understand that.

I think with Rondale Moore, for example, it’s important to note the distinction that the reason players his size haven’t succeeded is because the sample size is very low, and not because a great number of them have failed. There’s a difference there, at least to me.

Could be! Though I did note there have been 125 sub 5’9” receivers at the combine over the years, three of them have become fixtures in NFL offense, and none of them were under 5’8”.

That said, Moore has a lot of the same traits that have made Hill a successful outlier.

 

[PackFan2]

Teams regularly eliminate or significantly drop prospects because of off-field concerns, trying to avoid guys who will bust because they're knuckleheads, or just no committed.

100% I took one look at Randy Gregory's Twitter the year he was to be drafted (ALOT of weed references & pouting) and I was like nope.

 

[PikeBadger]

100% I took one look at Randy Gregory's Twitter the year he was to be drafted (ALOT of weed references & pouting) and I was like nope.

Perfect fit for Jerry Jones Cowboys. Lol.

 

[Dantés]

Brady Christensen and Alijah Vera-Tucker are both outliers if viewed as tackles due to arm length.

Christensen at 32.25" and Vera-Tucker at 32.13" are results that will almost certainly mean a move inside.

Good players. Good prospects. Almost certainly not tackles.

 

[PackFan2]

Good players. Good prospects. Almost certainly not tackles.

Idk man.. People said the same thing about Bakhtiari moving to guard... arm too short, not tall enough ETC. Hence outlier

 

[Dantés]

Idk man.. People said the same thing about Bakhtiari moving to guard... arm too short, not tall enough ETC. Hence outlier

Yeah, that's not correct.

Yes, some people projected him to move to guard.

But his arms are 34" long, which is totally normal for NFL tackles.

So his projection is not comparable to AVT or Christensen.

 

[Dantés]

Tutu Atwell is 155 lbs at 5087.

I scratch him from all draftable lists with great prejudice.

 

[Dantés]

A few other OL who have been debated tackle vs. guard are Jalen Mayfield (Michigan), Brenden James (Nebraska), and Jackson Carman (Clemson).

Arm length means that all three should probably be looked at as guards.

Mayfield: 32.63"
Carman: 32.5"
James: 32.75"

 

[tynimiller]

Tutu Atwell is 155 lbs at 5087.

I scratch him from all draftable lists with great prejudice.

Tutu's skillset is 2nd/3rd round worthy, Tutu the player I cannot even consider till mid to late Day 3 as a luxury dice role. Dude is just TINY