Jump to content
Sign in to follow this  
AlNFL19

Cornerback Class Statistical Projections

Recommended Posts

Here's the latest in my set of statistical projection models for the draft, using this year's class. The model takes into account several different factors and accordingly projects an average Approximate Value during the player's third and fourth NFL seasons. The model is built *on and for* prospects drafted in the first three rounds of the draft, using data from the 2006 to 2015 draft classes. It also projects a Bust Chance (% chance of failing to reach 5 AV) and Pro Bowl chance (chance of hitting or exceeding 10 AV), as well as a stat line based on the projected AV. 

The model includes:

  • Draft Position
  • College Statistics (INTs and PDs are not included because they tend to be driven by circumstance, making them fluctuate from year to year and thereby not be useful predictive tools)
  • NFL Combine Performance and Measurables
  • Speed Score
  • Waldo Explosive Power Formula

It's worth noting that the model is very conservative (no player has ever had a projection above 5 AV) because cornerbacks drafted don't tend to have a lot of NFL success (those drafted in Rounds 1-3 from 2006 to 2015 averaged around 2 AV in years 3 and 4). Nevertheless, the model's projections are correlated to real AV at 0.53, which is mediocre but a small step up from AV implied by draft position (0.48). 

For context, here's what a projection looks like, in the form of the best projection since 2006:

PATRICK PETERSON, LSU

P. Peterson, CB, LSU
Statistic Figure
Projected AV 4.73
Bust Chance 12.63%
PB Chance 19.25%

Projected Stat-Line: 47 tackles, 9 passes defensed, 2 INTs for 29 return yards

As I said, cornerback projections with the model tend to be conservative.

This year's projections are lower than most - it wasn't a fabulous class - and have some surprises, starting at number one:

 

1. SEAN BUNTING, CENTRAL MICHIGAN

S. Bunting, CB, Central Michigan
Statistic Figure
Projected AV 3.43
Bust Chance 60.24%
PB Chance 5.71%

Projected Stat Line: 43 tackles, 8 passes defensed, 1 INT for 22 return yards

Bunting was decent, but not a star, in every regard in the model, but that was good enough for first place in this year's class. His 1.51 10-yard split and his Explosive Power score both bested the group, helping him to a mediocre projection.

2. LONNIE JOHNSON JR., KENTUCKY

L. Johnson Jr., CB, Kentucky
Statistic Figure
Projected AV 3.31
Bust Chance 64.64%
PB Chance 4.46%

Projected Stat Line: 43 tackles, 8 passes defensed, 1 INT for 21 return yards

Johnson picked up steam late in the process as a bigger cornerback with long arms, but his speed testing at the Combine was below par, as is his projection.

3. TRAYVON MULLEN, CLEMSON

T. Mullen, CB, Clemson
Statistic Figure
Projected AV 3.26
Bust Chance 66.47%
PB Chance 3.93%

Projected Stat Line: 43 tackles, 8 passes defensed, 1 INT for 21 return yards

In terms of stats, Mullen's projection equals Lonnie Johnson's, but Mullen falls just short in projected AV. Mullen did lead the group in one statistic - tackles for loss, of which he had four.

4. JAMEL DEAN, AUBURN

J. Dean, CB, Auburn
Statistic Figure
Projected AV 3.25
Bust Chance 66.84%
PB Chance 3.83%

Projected Stat Line: 43 tackles, 8 passes defensed, 1 INT for 21 return yards

I'm sensing a theme with the projected stat lines with this group. Despite being drafted 94th overall, Dean's projection is third overall on account of his 4.3 speed.

5. ROCK YA-SIN, TEMPLE

R. Ya-Sin, CB, Temple
Statistic Figure
Projected AV 3.19
Bust Chance 69.03%
PB Chance 3.21%

Projected Stat Line: 42 tackles, 7 passes defensed, 1 INT for 20 return yards

Ya-Sin's 3.192 projection edges out first round pick Deandre Baker's, but obviously not by much. Ya-Sin, like Baker, turned in average Combine results that led to an average projection.

6. DEANDRE BAKER, GEORGIA

D. Baker, CB, Georgia
Statistic Figure
Projected AV 3.19
Bust Chance 69.03%
PB Chance 3.21%

Projected Stat Line: 42 tackles, 7 passes defensed, 1 INT for 20 return yards

Finally, the first round pick of the group. According to my model, Daniel Jones isn't the only first-round pick the Giants goofed. Baker's results were very average across the board, and his draft position wasn't enough to save him.

7. GREEDY WILLIAMS, LSU

G. Williams, CB, LSU
Statistic Figure
Projected AV 3.10
Bust Chance 72.33%
PB Chance 2.27%

Projected Stat Line: 42 tackles, 7 passes defensed, 1 INT for 20 return yards

For much of the early process, Greedy was the uncontested number one cornerback, but in the draft - and my model - he slipped. Williams showed good speed at the combine and decent athletic traits, but his middling draft position and on-the-shorter-side (31.5") arms docked his projection.

8. JOEJUAN WILLIAMS, VANDERBILT

J. Williams, CB, Vanderbilt
Statistic Figure
Projected AV 3.00
Bust Chance 75.99%
PB Chance 1.23%

Projected Stat Line: 42 tackles, 7 passes defensed, 1 INT for 19 return yards

I read an article the other day proclaiming Williams as the "Patriots' Richard Sherman." The model says one thing, loud and clear: slow your roll on that, SBNation. Williams put up decent explosion numbers and a bigger frame than many corners, but his 4.64 40 time cannot be ignored. Obviously, there is a chance for Williams to have success even without gamebreaking speed, it's just, according to the model, not something I would put money on.

9. BYRON MURPHY, WASHINGTON

B. Murphy, CB, Washington
Statistic Figure
Projected AV 2.99
Bust Chance 76.36%
PB Chance 1.12%

Projected Stat Line: 42 tackles, 7 passes defensed, 1 INT for 19 return yards

Murphy finished the draft season as the No. 1 corner on many analysts' rankings. Not this one. At a touch over 30", Murphy's arms are the shortest among those on the rankings, a fact that shouldn't be totally pushed aside even for a probable slot cornerback at the NFL level. Mostly, though, Murphy was very mediocre at the combine and on the stat sheet, and it shows here.

10. DAVID LONG, MICHIGAN

D. Long, CB, Michigan
Statistic Figure
Projected AV 2.81
Bust Chance 82.95%
PB Chance <1.00%

Projected Stat Line: 41 tackles, 7 passes defensed, 1 INT for 18 return yards

Despite some impressive numbers, like a 39.5" vertical jump, Long lands at tenth of eleven top-three-round corners. His mediocre on-field production (17 total tackles, 0 TFLs, etc.) didn't make him stand out at all.

11. JUSTIN LAYNE, MICHIGAN STATE

J. Layne, CB, Michigan State
Statistic Figure
Projected AV 2.56
Bust Chance 92.11%
PB Chance <1.00%

Projected Stat Line: 40 tackles, 7 passes defensed, 1 INT for 17 return yards

Michigan's Long might not have done very well, per the model, but Michigan fans can take comfort in that Long beat rival Michigan State's top offering to the draft at the position. With a 37.5" vert and a 134" broad jump, Layne had some impressive results, but the model would be willing to bet against his success. Layne's 4.5 40-time is made worse by his slight frame (190 lbs), and his draft position and statistics didn't help him out much.

That's all folks, feel free to peruse the NFL Draft forum for my threads on other positions.

 

@Awsi Dooger @HoboRocket @Hunter2_1 @VegasDan @deathstar @sryan66611

Just mentioning a few guys who have expressed interest in these things before. Just wanted to say a big thank you to you guys for asking questions and all and helping to motivate me to do more of these.

Share this post


Link to post
Share on other sites
Posted (edited)

Wow. What a bad class.

Edit - For context and my own fandom, how did Jaire Alexander and Josh Jackson fare last year?

Edited by deathstar

Share this post


Link to post
Share on other sites
Posted (edited)
3 hours ago, deathstar said:

Wow. What a bad class.

Edit - For context and my own fandom, how did Jaire Alexander and Josh Jackson fare last year?

Their projections are conservative, like all of them (see above: Patrick Peterson's record projection was just a 4.73).

Alexander and Jackson have very similar builds (5'10", 196 lbs) but Alexander was more productive on the field and tested better at the Combine, so he's quite a ways ahead in projection.

J. Alexander, CB, Louisville
Statistic Figure
Projected AV 3.8
Bust Chance 46.69%
PB Chance 9.56%

Projected Stat Line: 44 tackles, 8 passes defensed, 2 interceptions for 24 return yards

J. Jackson, CB, Iowa
Statistic Figure
Projected AV 2.99
Bust Chance 76.36%
PB Chance 1.12%

Projected Stat Line: 42 tackles, 7 passes defensed, 1 interception for 19 return yards

Their stat line projections are similar, and with those above as well, because they're average stat lines built around the projected AV scores, which themselves aren't very far off from each other.

Edited by AlNFL19

Share this post


Link to post
Share on other sites

I know that David Long was a darling for a lot of draftniks and people here due to his insane athleticism. Actually reminded me a lot of how Jaire Alexander was last year in off-season workouts with his smooth movements and ability to flip his hips. The dude has good deep speed (4.45), but his quickness and acceleration are insane (3.97 short shuttle and 6.45 3-cone were the best among ANYONE in this class). That's on top of a top 5 vertical jump and top 5 bench press in the DB class. He didn't put up great gross number compilation stats, but his coverage numbers are fantastic when you consider that, on his career, he allowed the lowest passer rating in the Pro Football Focus era with a rating of 11.9 and only 1 TD pass allowed in his career. He only allowed 18 receptions on 60 targets and just 2 first downs in 2018. He was an amazing press man corner that would just lockdown receivers so that he wasn't really targeted. That makes his small gross number compilation stats a lot more impressive. 

Share this post


Link to post
Share on other sites

Crazy how low those AVs are for CBs on average.  What was Marshon Lattimore's class like?  There were a few stud CBs from that class.

Share this post


Link to post
Share on other sites
3 hours ago, Raves said:

Crazy how low those AVs are for CBs on average.  What was Marshon Lattimore's class like?  There were a few stud CBs from that class.

Here's the 2017 class:

2017 NFL Cornerback Draft Class
Round Selection Name College Weighted Avg. AV Bust Chance PB Chance Proj. Tackles Proj. INTs Proj. Ret. Yds Proj. PDs
1 11 M. Lattimore Ohio State 3.86 44.49% 10.18% 44 2 24 8
1 16 M. Humphrey Alabama 3.75 48.52% 9.04% 44 2 24 8
1 24 G. Conley Ohio State 3.55 55.85% 6.96% 44 2 22 8
2 33 K. King Washington 3.5 57.68% 6.43% 43 2 22 8
1 18 A. Jackson USC 3.49 58.05% 6.33% 43 2 22 8
1 27 T. White LSU 3.4 61.34% 5.39% 43 1 22 8
2 46 Q. Wilson Florida 3.29 65.37% 4.25% 43 1 21 8
3 81 F. Moreau UCLA 3.25 66.84% 3.83% 43 1 21 8
2 60 C. Awuzie Colorado 3.19 69.03% 3.21% 42 1 20 7
3 66 A. Witherspoon Colorado 2.99 76.36% 1.12% 42 1 19 7
2 43 S. Jones Washington 2.94 78.19% <1.00% 42 1 19 7
3 90 S. Griffin UCF 2.84 81.85% <1.00% 41 1 18 7
2 53 T. Tabor Florida 2.76 84.79% <1.00% 41 1 18 7
3 97 C. Tankersley Clemson 2.72 86.25% <1.00% 41 1 18 7
3 99 R. Douglas West Virginia 2.43 96.87% <1.00% 40 1 16 6
3 92 J. Lewis Michigan 2.28 >99.9% <1.00% 40 1 15 6
3 94 C. Sutton Tennessee 2.14 >99.9% <1.00% 39 1 14 6

Lattimore falls just below Denzel Ward's 3.90.

The projections are always going to be low for CBs because CB performance tends to be low. From the sample the model is built on - cornerbacks drafted in Rounds 1-3 from 2006 to 2015 - the mean AV is 3.12, and that's including Darrelle Revis averaging a quarterback-like 17.5 AV in his third and fourth years. Take Revis out, and it drops to 2.99. Patrick Peterson is the only guy to ever come close to 5 AV (4.73), at 0.54 ahead of every other corner. A big part of that is his insane athletic profile: Peterson ran a 4.31 at 219 pounds at the Combine. 

Share this post


Link to post
Share on other sites

So, this might be a dumb question, but if the statistical projection for cornerbacks is so conservative to the point of labeling almost every prospect a likely bust, can't you change how it evaluates for better results?

Share this post


Link to post
Share on other sites
7 minutes ago, Daniel said:

So, this might be a dumb question, but if the statistical projection for cornerbacks is so conservative to the point of labeling almost every prospect a likely bust, can't you change how it evaluates for better results?

I mean, if it’s in line with what’s happened in the past, why change it to make more guys come out as not busts if it’s not very likely they do? That’s my line of thinking, anyway. The majority of drafted cornerbacks will be busts, so projecting most players as likely to do so makes sense. 

Share this post


Link to post
Share on other sites

good stuff @AlNFL19

 

Its nice to see the guys Bucs grabbed are highly rated but it just seems like a bad draft class in general for the CB's...   SMB (Sean Murphy Bunting) has been pretty dang good in camp so far he's coming along nicely and will probably have a nice role in the defense once the regular season starts.

Share this post


Link to post
Share on other sites
On 7/31/2019 at 3:38 PM, AlNFL19 said:

I mean, if it’s in line with what’s happened in the past, why change it to make more guys come out as not busts if it’s not very likely they do? That’s my line of thinking, anyway. The majority of drafted cornerbacks will be busts, so projecting most players as likely to do so makes sense. 

I appreciate that you don't do it that way. Conventional wisdom regarding cornerbacks has been wrong, and particularly during the spotlighted decade. It was brutally wrong here and elsewhere. If someone here had been mocking the cornerback criteria as way out of line they would have been steadily ridiculed and it would have taken a heck of a long time for that criticism to soften. I'm still not sure it would have happened. 

Reading the Seahawks Draft Blog from Rob Staton over the past 6-7 years has really aided my perspective on cornerbacks. Those Seattle bloggers have been way above the rest of the league in terms of identifying various traits and test scores vital for each position. Just because the Seahawks started making some strange choices didn't mean the content was flawed. Many years ago it reached the point I looked for the 32 inch arms as a dividing line, not much different than 9 inch hands at quarterback. However, the percentages are vastly different. Not many top quarterback prospects have sub-9 inch hands. In contrast, the majority at cornerback are below 32 inch arms. 

Mel Kiper was mostly jumping in line when he started ripping the running back position as overstated, and not worth first round inclusion. Meanwhile if Kiper had avoided that cliche and instead shifted focus to cornerbacks as overdrafted, with better ones available late, he would have been first in line and eventually noted as seer.

There is some type of stat from Rob Staton regarding multiple All Pro nods and 32 inch arms. I am sorry that I can't remember it specifically. I think it is 20 years or more since a cornerback with sub-32 inch arms was selected as All Pro more than once. The shorter armed guys can jump up and look excellent for a while, but it is mostly hype and eventually they settle. As a Dolphins fan I keep that in mind with Xavien Howard and his 31 1/4 arms. It is chic to gush over Howard right now. I am not falling for it. The guy was not an elite recruit. He was not an elite draft pick. He doesn't have ideal dimensions. Good player but I think he will level out and mostly disappoint, over the length of this contract.

Share this post


Link to post
Share on other sites
Posted (edited)
57 minutes ago, Awsi Dooger said:

I appreciate that you don't do it that way. Conventional wisdom regarding cornerbacks has been wrong, and particularly during the spotlighted decade. It was brutally wrong here and elsewhere. If someone here had been mocking the cornerback criteria as way out of line they would have been steadily ridiculed and it would have taken a heck of a long time for that criticism to soften. I'm still not sure it would have happened. 

Reading the Seahawks Draft Blog from Rob Staton over the past 6-7 years has really aided my perspective on cornerbacks. Those Seattle bloggers have been way above the rest of the league in terms of identifying various traits and test scores vital for each position. Just because the Seahawks started making some strange choices didn't mean the content was flawed. Many years ago it reached the point I looked for the 32 inch arms as a dividing line, not much different than 9 inch hands at quarterback. However, the percentages are vastly different. Not many top quarterback prospects have sub-9 inch hands. In contrast, the majority at cornerback are below 32 inch arms. 

Mel Kiper was mostly jumping in line when he started ripping the running back position as overstated, and not worth first round inclusion. Meanwhile if Kiper had avoided that cliche and instead shifted focus to cornerbacks as overdrafted, with better ones available late, he would have been first in line and eventually noted as seer.

There is some type of stat from Rob Staton regarding multiple All Pro nods and 32 inch arms. I am sorry that I can't remember it specifically. I think it is 20 years or more since a cornerback with sub-32 inch arms was selected as All Pro more than once. The shorter armed guys can jump up and look excellent for a while, but it is mostly hype and eventually they settle. As a Dolphins fan I keep that in mind with Xavien Howard and his 31 1/4 arms. It is chic to gush over Howard right now. I am not falling for it. The guy was not an elite recruit. He was not an elite draft pick. He doesn't have ideal dimensions. Good player but I think he will level out and mostly disappoint, over the length of this contract.

So you think the be all and end all as a CB is arm length.  32" arm length?  Denzel Ward does not have 32" arm length, neither does Chris Harris Jr and they both just made the Pro bowl and Harris has been on an All Pro team 3 times and Ward will be plenty more also.  

 

Antoine Winfield will be in the Hall of Fame shortly and he was quite short with short arms, Ty Law was just elected to the Hall of Fame not to mention and 5-11 under 32" arms I am sure.  

 

Donte Jackson, Marshon Lattimore, Jourdan Lewis, Vernon Hargreaves, Bradley Roby, Kendall Fuller, Avonte Maddox, Ronald Darby and potentially Byron Murphy, Mike Hughes, Jaire Alexander as well are damn good defensive backs, some really good and they do not have 32" arms.

 

 

The Seattle Seahawks are rare in the fact they can find guys who were underrated and make them into big time CBs.  Best example was Tre Flowers a S at Ok State and really underachieved but they put him in a position to succeed at corner and last year he did great.  Will see if he can keep that up and for how long.  Others would be Brandon Browner a guy who had a ton of talent but failed in other systems, same with Byron Maxwell.  They know how to use big corners and get them to play well in that system, helps having a Hall of Fame safety back there Earl Thomas no doubt for both Browner and Maxwell.  The best is Richard Sherman of course, they used his past WR skills and he did great in that system as a long corner.  

 

Are bigger corners sometimes over drafted, sure take Ras I Dowling, Stanley Jean Baptiste, Jonathan Banks, Jalen Collins, Eric Rowe, Chris Cook, Mike Rumph etc.  But that happens with a lot of positions.  With how the game is played now small corners are going to be very needed to cover the slot and guys like Donte Jackson, Byron Murphy etc will do very good things and be very important for defenses.  So what about their arm length.  

Edited by Ozzy

Share this post


Link to post
Share on other sites
1 hour ago, Ozzy said:

Are bigger corners sometimes over drafted, sure take Ras I Dowling, Stanley Jean Baptiste, Jonathan Banks, Jalen Collins, Eric Rowe, Chris Cook, Mike Rumph etc.  But that happens with a lot of positions.  With how the game is played now small corners are going to be very needed to cover the slot and guys like Donte Jackson, Byron Murphy etc will do very good things and be very important for defenses.  So what about their arm length.  

Arm length isn't an end-all-be-all for the cornerback position - nothing is, really - but it is a decent predictor in comparison with other statistics and measurables (correlating to AV in the sample I used a little better than 40 time). I think it's a helpful measurable when taken in context - if a cornerback doesn't have the speed to keep up with receivers, he will be better served by long arms that will make up that half-step he might lose in a foot race. But as you explained, there are plenty of examples of cornerbacks who don't have long arms - generally in the same mold of a smaller, quicker corner who has the speed to keep up.

Share this post


Link to post
Share on other sites
2 minutes ago, AlNFL19 said:

Arm length isn't an end-all-be-all for the cornerback position - nothing is, really - but it is a decent predictor in comparison with other statistics and measurables (correlating to AV in the sample I used a little better than 40 time). I think it's a helpful measurable when taken in context - if a cornerback doesn't have the speed to keep up with receivers, he will be better served by long arms that will make up that half-step he might lose in a foot race. But as you explained, there are plenty of examples of cornerbacks who don't have long arms - generally in the same mold of a smaller, quicker corner who has the speed to keep up.

Seattle is a prime example of a team that knows how to use long corners and can make them successful especially when they were late round picks.  No other team has a track record like that to be honest of multiple tall long corners being successful that were late round picks. Sure a bigger corner is usually better but to say 32" arms or nothing is just stupid and not correct.  

 

Champ Bailey and Charles Woodson as college prospects are arguably two of the best corners to ever play the position.  Their upside to me was greatly because of their ball skills from being a wide receiver.  Being able to react to the football in the air and make plays.  Also helps both were great competitors and tacklers no question.  Sure it helps they were tall but their ball skills and how they played were more important than their height.  Oh and Champ Bailey also did not have 32" arms and was only 6-0.  

 

Oh and Terence Newman had a great career as a smaller CB, not to mention Darrell Green and Deion Sanders obviously.  Jonathan Joseph is another one still doing it at a decent level.  Cortland Finnegan was a great player for a little while I thought.  Vontae Davis was a monster at corner and physically amazing I thought, another shorter dude shorter arm length.

 

 

Helps that some bigger corners or just corners in general who cannot quite make it at CB can move to safety and still be very good especially Malcolm Jenkins, Devin McCourty, Marlin Jackson, Tyrann Mathieu, Micah Hyde and to a lesser extent Chris Culliver and  Eugene Wilson.

Share this post


Link to post
Share on other sites
Posted (edited)

Great work, @AINFL19 love this kind of analysis.  And good luck on your predictions on stays..  If you hit these stats within 20% i would expect:

18.30% chance NFL scouting Dpt unpaid intern offer

7.40% chance NFL draft analyst job offer

1.30% asst GM/ job offer

0.01% chance NFL GM job offer

Edited by VegasDan

Share this post


Link to post
Share on other sites
Posted (edited)

I have many questions. You don’t use college interceptions or PDs as an input variable, but you’re predicting interceptions and PDs? What college statistics exactly are you using? What variables are you regressing onto when training your model? What constitutes a bust or a success? What is the correlation coefficient?

It seems to me that your model is heavily weighted on physical attributes. Notice your top 5 guys are 6 feet tall, near 200 pounds, and run 4.5 or faster 40s at the combine. If it were that simple, the draft wouldn’t be such a crap shoot. Furthermore, your features are redundant and not independent. Waldo’s formula, speed score, and combine numbers all boil down to the same features - combine numbers. 

DeAndre Baker isn’t the biggest or fastest corner, but he won the Jim Thorpe award for a reason - he’s a darn good corner on the field. The reports from camp have been nothing but positive and he’s already listed as a starter. 

I’m a big fan of statistics and models since I work in machine learning. But my confidence in your methodology is low.

Edited by Gmen

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now
Sign in to follow this  



×