FanPost

The Best QB vs The Best Defensive Player: A 15-Year Draft Analysis

It's official! Your Houston Texans will be drafting first overall in the 2014 Player Selection Meeting.

The biggest discussion revolving around this draft has been about the two players likely to be selected by the Texans: Teddy Bridgewater or Jadeveon Clowney. In one corner, you have the head-and-shoulders best QB available, a position that has proven repeatedly to be a make-or-break position for a franchise. Some Texans fans MIGHT be aware of how debilitating it is for a playoff-caliber team to suffer from poor QB play. (For that matter, watching the Colts for most of the past 20 years shows you how much great QB play can cover up a lot of weaknesses).

That said - and keep in mind that I am firmly on Team Teddy - the best player available is Clowney. The knocks against him aside, the odds of him being anything less than a good-to-great NFL pass rusher are about as low as the odds of me suiting up for an NFL team. Most likely, Clowney will be a perennial Pro Bowler who will easily eclipse 100 sacks in his career.

So the question is simple: draft the best player at the most important position, where we have a need? Or do we draft the best available player regardless of position?

The answer, to me at least, requires the question to shift. See, Lord knows we are drafting a QB. Even if we pass on Teddy, that just means we are waiting until Round 2 or 3. So the question isn't Teddy vs. Clowney; the question is which is better: QB, then defense, or defense THEN QB?

I decided to analyze four players from each of the last 15 drafts:
-The first QB selected
-The first QB selected AFTER Round 1 ends
-The first defensive player selected
-The first player AT THE ORIGINAL DEFENDER'S POSITION after Round 1

I chose to use the same position to make the comparison more apples-to-apples. Plus, it makes sense that, if the team is interested in a stud DE, chances are the team has a glaring need there and would draft that position later.

The players were analyzed using advancednflstats.com and their WPA and EPA stats. The former calculates how many wins a player adds (or subtracts) from his team, while the latter calculates how many points he helps his team produce. These stats, which the website tracks for both offensive and defensive players, are my basis for comparison. For each player, I have listed:
-Their WPA for their rookie year
-Their EPA for their rookie year
-Their cumulative WPA for their first 3 seasons
-Their cumulative EPA for their first 3 seasons

I calculated 1-year and 3-year increments because I want this draft to launch us back into immediate playoff contention, but it can't be overlooked that I want long term success for the franchise as well.

Here's the cast. I have their stats, as outlined above, beside their results. Obviously, players from the 2012 and 2013 draft class have incomplete "3-year results".

First-Round Quarterbacks

E.J. Manuel
0.03 / -27.0 / 0.03 / -27.0

Andrew Luck
4.50 / 107.6 / 6.98 / 176.7

Cam Newton
1.91 / 106.3 / 7.46 / 316.0

Sam Bradford
0.46 / -10.2 / 0.53 / -61.1

Matthew Stafford
0.71 / -47.5 / 4.44 / 64.3

Matt Ryan
3.82 / 89.7 / 10.69 / 266.3

JaMarcus Russell
0 / 0 / -0.99 / -117.6

Vince Young
0.81 / 10.9 / 0.96 / 7.6

Alex Smith
-1.40 / -79.3 / -3.45 / -160.6

Eli Manning
-.06 / -24.9 / 2.41 / 35.2

Carson Palmer
0 / 0 / 3.86 / 96.7

David Carr
-2.72 / -101.2 / -1.08 / -73.5

Michael Vick
-0.52 / -19.2 / 0.72 / 62.1

Chad Pennington
0 / 0 / 2.84 / 98.6

Tim Couch
-.06 / -40.4 / -1.68 / -115.4

So, basically, the stats indicate that, for example, Matt Ryan was individually responsible for making the Falcons about 4 wins better in his rookie season, and 10 wins better as a franchise the first 3 years. On the other hand, David Carr cost his team 101 points his rookie year (a 16-game average of nearly a whole touchdown), but improved enough to only cost the Texans 73.5 points his first 3 years total (a per-game average of only about 1.5).

Second-Round Quarterbacks

Geno Smith
-0.81 / -55.4 / -.081 / -55.4

Brock Osweiler
0 / 0 / 0 / 0

Andy Dalton
-.02 / 29.0 / 1.18 / 93.4

Jimmy Clausen
-1.39 / -83.7 / -1.39 / -83.7

Pat White
0 / 0 / 0 / 0

Brian Brohm
0 / 0 / 0 / 0

John Beck
0 / 0 / 0 / 0

Kellen Clemens
0 / 0 / -1.19 / -25.7

Charlie Frye
-0.14 / -23.1 / -1.81 / -71.6

Matt Schaub
0 / 0 / 0 / 0

Dave Ragone
0 / 0 / 0 / 0

Josh McCown
0 / 0 / -1.31 / -21.7

Drew Brees
0 / 0 / -.06 / -26.1

Giovanni Carmazzi
0 / 0 / 0 / 0

Shaun King
0.03/-7.4/-.32/24.9

Ten out of the last 15 QB's drafted after Round 1 did not play (or, to be technical, did not accrue significant stats) their rookie season. Is that an indicator that QB's graded outside of the first round just aren't NFL ready? Maybe. Also, I'll get to the statistical comparison in a minute, but does anyone else notice that only ONE OUT OF FIFTEEN post-first QBs had a long-term positive affect on his team's win production?

Best Defensive Player

Dion Jordan - DE
0 / 0 / 0 / 0

Morris Claiborne - CB
0.91 / 29.6 / 1.23 / 43.3

Von Miller - OLB
1.95 / 53.3 / 4.69 / 174.6

Ndamukong Suh - DT
1.37 / 48.3 / 3.57 / 114.9

Tyson Jackson - DE
0.41 / 9.4 / 0.41 / 9.4

Chris Long - DE
0.65 / 22.9 / 2.64 / 88.6

Gaines Adams - DE
0.79 / 31.2 / 2.50 / 75.0

Mario Williams - DE
0.77 / 31.7 / 3.14 / 129.8

Adam "PacMan" Jones - CB
0.67 / 18.1 / 1.90 / 57.4

Sean Taylor - S
1.06 / 38.5 / 2.34 / 77.8

DeWayne Robertson - DT
0.45 / 15.8 / 1.95 / 60.2

Julius Peppers - DE
1.62 / 46.2 / 4.81 / 144.5

Gerard Warren - DT
1.20 / 38.8 / 3.13 / 95.2

Courtney Brown - DE
1.76 / 44.9 / 2.63 / 71.2

Champ Bailey - CB
1.10 / 31.3 / 2.99 / 94.0

For some reason, defensive players do not have negative scores on WPA or EPA. for the purpose of this analysis, I'm willing to use that to numerically quantify the importance of the QB position, which we are not debating. Most people in favor of drafting Clowney agree that QB is the most important position; they just don't think that would justify an inferior prospect.

Second-Round Defensive "Replacements"

Tank Carradine - DE
0 / 0 / 0 / 0

Janoris Jenkins - CB
1.56/47.9/1.87/65.7

Jabaal Sheard - OLB
1.41/47.8/2.25/78.0

Brian Price - DT
0 / 0 / 0 / 0

Everette Brown - DE
0 / 0 / 0 / 0

Phillip Merling - DE
0 / 0 / 0.68 / 19.4

LaMarr Woodley - DE
0 / 0 / 3.81 / 109.5

Darryl Tapp - DE
0 / 0 / 3.01 / 102.9

Stanford Routt - CB
0.15 / 4.0 / 0.56 / 20.6

Bob Sanders - S
0.05 / 12.5 / 0.87 / 47.0

Anthony Adams - DT
0 / 0 / 1.09 / 41.0

Kalimba Edwards - DE
0.93 / 28.2 / 0.93 / 28.2

Kris Jenkins - DT
0.42 / 12.1 / 2.81 / 77.8

Darren Howard - DE
1.18 / 44.1 / 3.42 / 110.8

Charles Fisher - CB
0 / 0 / 0 / 0

Three of the fifteen here got over the 3.0 hump on cumulative WPA (versus 6 first-rounders, because Champ's 2.99 totally counts) but only 3 (versus 7 first-rounders) had a bigger than 1.0 impact on their rookie season.

Statistical Comparison

Two things are obvious, though we have re-proven it: First rounders are clearly better than second rounders. In both fields of players, the guys drafted earlier were more likely to make an immediate as well as a lasting impact in the positive direction.

My first effort to produce meaningful data was to simply have the pairs of players go head-to-head each year. (For example, Tim Couch + Charles Fisher vs. Shaun King + Champ Bailey). The results were inconclusive. Of the fifteen years, eight were in favor of drafting a QB early, 5 favored drafting defense first, and two were a tie (to include the incredibly incomplete 2013 class). A slight lean towards drafting Teddy, but not the sure thing for either side I was hoping for.

So, I went ahead and took the four categories of players and wrote down their statistical range as well as calculating their averages for each stat.

Early QB's
Rookie WPA(Wins): -2.72 - 4.50
Rookie EPA(Points): -101.2 - 107.6
3-Year WPA: -3.45 - 10.69
3-Year EPA: -160.6 - 316.0

Average:
0.49 / -1.03 / 2.41 / 40.6

2nd-Rd QB's
Rookie WPA: -1.39 - 0.03
Rookie EPA: -83.7 - 29.0
3-Year WPA: -1.81 - 1.18
3-Year EPA: -83.7 - 93.4

Average: -0.17 / -10.0 / 0.84 / -11.9

Now let's take a look at what happens if we compare these two pools of stats. In other words, what is the difference between first-round and second-round QB's from a range/average perspective?

QB1's - QB2's =
Rookie WPA: -1.33 - 4.47
Rookie EPA: -17.5 - 78.6
3-Year WPA: -1.64 - 9.51
3-Year EPA: -76.9 - 223.2

Average: 0.66 / 8.97 / 1.57 / 52.5

So, a bust at QB is greatly amplified if we draft one early; That first number indicates that the Texans fielding the worst-case scenario of #1 pick at QB would be a game worse than trotting out a worst-case 2nd-round QB. But, on the other side of the coin, the best possible 2nd-rounder is still 4.47 games worse than a home run at #1.

Let's look at the defenders:

Early Defensive Players
Rookie WPA: 0 - 1.95
Rookie EPA: 0 - 53.3
3-Year WPA: 0.41 - 4.81
3-Year EPA: 9.4 - 174.6

Averages: 0.98 / 30.6 / 2.71 / 91.1

Worth noting: ON AVERAGE, players like Clowney single-handedly make their team consistently a game better. That's with Clowney being on par with potential, but not exceeding it.

"Replacement" Defensive Players
Rookie WPA: 0 - 1.56
Rookie EPA: 0 - 47.9
3-Year WPA: 0 - 3.81
3-Year EPA: 0 - 110.8

Averages: 0.38 / 13.1 / 1.52 / 50.1

The numbers are evident without further detail here: a 2nd-round pick can be almost as effective as a 1st-rounder in the best-case, but the later picks come with a higher rate of being ordinary.

The Final Statistical Comparison

At this point, I feel the following are conclusive:
-a #1 pick at QB has the single-highest chance of being a grand-slam home run
-a failed #1 pick at QB is the worst possible thing that can happen to a franchise, within the realm of this study
-a 2nd-round pick at QB is very unlikely to produce an immediate or short-term player that will effectively increase our win/point output.
-a #1 pick of a defensive stud is the closest thing to a sure-fire improvement that a franchise can make, within the realm of this study
-Whereas a later QB is worse in their potential output, a later defender has just as high a ceiling, but comes with a smaller likelihood of achieving significant output.

Let's look at one final statistical analysis. I took the ranges and averages, and constructed them as so:

(#1QBs + #2 DEF) - (#2 QBs + #1 DEF)

It's the question I've been asking all post, and here is the statistical result. Remember, a negative result indicates a result in favor of waiting to draft a QB.

Rookie WPA: -1.43 - 4.08
Rookie EPA: -17.5 - 73.2
3-Year WPA: -2.05 - 8.51
3-Year EPA: -86.3 - 158.8

Averages: 0.06 / -8.50 / 0.38 / 11.5

The averages result indicates an unsatisfying result: On average, it turns out roughly the same.

However, the range of output is also telling in a different way. What I have taken from this output is simple: Drafting a QB at #1 is a high-risk/high-reward decision. It would be much safer (and more likely to produce a net-positive result) to draft Clowney #1 and take Tajh Boyd or whomever is Best QB available with the #33 pick. However, if Bridgewater DOES turn out to be more Newton/Luck than Carr/Couch, he will have a greater benefit to the Texans than Clowney/Boyd can ever be expected to achieve.

I do not think these results indicate an obvious option, and I think it comes down to whether you think that a conservative draft strategy is better for the franchise than an aggressive one.

X
Log In Sign Up

forgot?
Log In Sign Up

Forgot password?

We'll email you a reset link.

If you signed up using a 3rd party account like Facebook or Twitter, please login with it instead.

Forgot password?

Try another email?

Almost done,

Join Battle Red Blog

You must be a member of Battle Red Blog to participate.

We have our own Community Guidelines at Battle Red Blog. You should read them.

Join Battle Red Blog

You must be a member of Battle Red Blog to participate.

We have our own Community Guidelines at Battle Red Blog. You should read them.

Spinner

Authenticating

Great!

Choose an available username to complete sign up.

In order to provide our users with a better overall experience, we ask for more information from Facebook when using it to login so that we can learn more about our audience and provide you with the best possible experience. We do not store specific user data and the sharing of it is not required to login with Facebook.

tracking_pixel_9341_tracker