FanPost

Quantifying Success Probability: Quarterbacks

In competitive sports there has been a documented use of advanced statistics to try to gain a competitive advantage. The most documented statistic was made famous in the Michael Lewis book Moneyball in which the Oakland Athletics used advanced statistics to determine undervalued players. While Oakland, and baseball for that matter, no longer value the same stats they once did, there continues to be different measures being developed from new models each year. Nice right? Now for the relevant part: Baseball is not the only game that uses advanced statistics.

Oakland revealed its success and revealed a core competency with the book Moneyball. In 2003 after their "misfit" players tied the Yankees for most wins with 103. Although both teams won 103 games, they had remarkably different approaches. The Yankees found players who passed the eye test, had excellent well known stats such as average, home runs, etc, and signed them to lucrative contracts. Oakland saw an undervalued stat: On Base Percentage and saw that speed and defense were over valued. The Yankees spent $133M while the A's spent $41M.

Stats in football are not well known to the public, but there are certainly teams who appear to be taking advantage of statistics. San Francisco certainly appears to be one of those teams as does Seattle. With John Idzik now in charge of the Jets front office, is it possible the Jets will also be using statistics to find value. Link to an interested article that suggested the Alex Smith trade was aided by stats can be found here: http://www.newrepublic.com/article/112300/football-sabermetrics-how-stats-could-transform-nfl

Now, I'll be the first to admit, I am not a statistician. My econometrics class was one of the hardest classes I may have taken. In all advanced stats models you are trying to optimize the determining factors that make up a high Coefficient of determination. In laments terms, we need to identify what factors make up a high probability of success. In stats, it is necessary to minimize biases. When the common amateur scout views a video, they are swayed by several types of biases: Play type, something different, team, surrounding talent, coaches, height, weight, measurables, conferences, etc. The most common way to reduce bias is to evaluate a set of criteria that is critical to success and measure the accuracy. Then have a high sample size evaluate a player, and enter quantitative numbers into your regression model.

I do not have the software to determine actuals, but if I were trying to build a regression model for quarterbacks, I would probably bet on the follow factors determining success.

Trait

Impacts

Qualitative/Quantitative

Correlation Factor

Current

Projected

Height

Completion percentage, Field Vision

Quantitative

0.04

BMI

Durability, Strength

Quantitative

0.03

Hand Size

Completion percentage, Ball Security

Quantitative

0.06

Acceleration

Mobility

Quantitative

0.02

Velocity

Arm Strength

Quantitative

0.04

Release Time

Completion percentage, Ball Security

Quantitative

0.08

40 Yard Dash

Mobility

Quantitative

0.005

Number of starts

Experience

Quantitative

0.04

Wonderlic

Field Vision, Growth Potential

Quantitative

0.07

Work Ethic/Dedication

Growth Potential

Qualitative

0.08

Completion Percentage adjusted for drops under 5 yards

Completion percentage

Quantitative

0.05

Completion Percentage adjusted for drops 5-15 yards

Completion percentage

Quantitative

0.06

Completion Percentage adjusted for drops 16+ yards

Completion percentage

Quantitative

0.04

Number of total throws under 5 yards

Experience

Quantitative

0.02

Number of total throws 5-15 yards

Experience

Quantitative

0.02

Number of total throws over 15 yards

Experience

Quantitative

0.02

Footwork

Arm Strength, Completion percentage

Qualitative

0.05

Conference

Experience

Qualitative

0.015

Leadership

Growth Potential

Qualitative

0.07

Maturity

Growth Potential

Qualitative

0.05

Pocket Presence

Field Vision

Qualitative

0.04

Total Plays out because of injury/Durability

Durability

Quantitative

0.07

Correlation factor adds up to .97 in my estimates leaving 3% undefined. Qualitative estimates generally have biases which is why more diverse opinions may reduce biases.

That would be my general scope for identifying success of finding a QB in the draft and not getting caught up in selected available film and hype.

Another statistical items to rate QBs is Lewin Career Forecast. It takes 7 aspects into effect.

  • Career college games started
  • Career completion rate. Because of recent rises in completion rate across college football, this is a logarithmic variable, so that as a quarterback's completion percentage goes down, the penalty for low completion percentage gets gradually larger.
  • Difference between the quarterback's BMI and 28.0. This creates a small penalty for quarterbacks who don't exactly conform to the "ideal quarterback size."
  • For quarterbacks who come out as seniors, the difference in NCAA passer rating between their junior and senior seasons. (For quarterbacks who come out as juniors or redshirt sophomores, this variable is always 5.0, which is the average increase for the seniors in data set.)
  • A binary variable that penalizes quarterbacks who don't play for a team in a BCS-qualifying conference.
  • Run-pass ratio in the quarterback's final college season.
  • Total rushing yards in the quarterback's final college season.

This forecast showed Geno was the best prospect in the last draft class. Russell Wilson has amongst the highest projections as did RG3 and Phillip Rivers. *From Football Outsiders http://www.footballoutsiders.com/stat-analysis/2013/lewin-career-forecast-2013 *

The most basic statistical test is the 26-27-60 Test. It suggests a QB will have a much higher success rate in the NFL if they score:

  • at least 26 on the Wonderlic
  • at least 27 games started
  • at least 60% completion rate

FYI... Geno did not pass. He scored a 24 on the Wonderlic.

Do you believe in such statistical models? Do you prefer one over the other?



SB Nation Featured Video
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,

By becoming a registered user, you are also agreeing to our Terms and confirming that you have read our Privacy Policy.

Join Gang Green Nation

You must be a member of Gang Green Nation to participate.

We have our own Community Guidelines at Gang Green Nation. You should read them.

Join Gang Green Nation

You must be a member of Gang Green Nation to participate.

We have our own Community Guidelines at Gang Green Nation. You should read them.

Spinner.vc97ec6e

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