Writers of Pro Football Prospectus 2008

29 Mar 2011

ESPN: Playmaker Score 2.0

This ESPN Insider column introduces Playmaker Score 2.0, a more accurate version of our system for projecting college wide receivers. The new Playmaker Score does more to account for players who go deep with a high yards per pass figure, and it adjusts based on the historical record of receivers coming out of each conference. The resulting system has A.J. Green as the top receiver in this year's draft, but wonders why Julio Jones had so few touchdowns at Alabama.

For those of you who don't have Insider, expect to see a longer version of this article in Football Outsiders Almanac 2011.

Posted by: Vincent Verhei on 29 Mar 2011

10 comments, Last at 12 Apr 2012, 2:41pm by DJensen

Comments

1
by ml19x (not verified) :: Tue, 03/29/2011 - 6:45pm

Any chance we can get a peek at the formula for Playmaker 2.0?

3
by dafrk3in :: Wed, 03/30/2011 - 1:32pm

My guess is that they look at the observed relationship between college performance and observables related to the player's team's style of play and conference, run a regression to figure out the weights to put on various factors, and use these weights to predict the performance of this year's crop of WR candidates.

5
by Vincent Verhei :: Wed, 03/30/2011 - 5:35pm

Any chance we can get a peek at the formula for Playmaker 2.0?

Awesome. The formula was in the original version of the story, my ESPN editor asked me to take it out, and the very first FO comment is asking to put it back in.

The formula starts by multiplying four factors together:

* Career yards per game.
* Career yards per catch.
* Career touchdowns.
* Team yards per attempt.

Divide that product by team passes per game, and divide THAT product by 20. Then multiply the result for a conference factor to adjust for quality of teammates/opposition/weather:

ACC 2.087
Big Ten/ND 1.463
SEC 1.331
NAQ 1.088
Big 12 0.853
Pac-10 0.807
Big East 0.736

8
by cozmark :: Tue, 05/17/2011 - 4:10pm

deleted by author

2
by Aaron Brooks' Good Twin (not verified) :: Wed, 03/30/2011 - 9:46am

Because Alabama has an overwhelming offensive line and a stable of excellent power backs? They just Smithed it into the end zone whenever they got close.

7
by Formersd (not verified) :: Mon, 04/04/2011 - 11:30am

Still doesn't explain why Jones only caught 17 of the 52 passing TDs they had while he was there. The way I'm reading the results, that's the factor the formula dinged him for more than anything else. Confirms my concern with Jones, he's an amazing athlete, but does that mean he's a great receiver?

Interestingly, Hankerson was on my radar as someone I needed to look into more, glad to see another reason to take a closer look.

4
by dafrk3in :: Wed, 03/30/2011 - 1:34pm

Have you made any attempt to factor in the team that the WR plays for to get team-specific effects on NFL performance?

6
by ml19x (not verified) :: Fri, 04/01/2011 - 4:53pm

Thanks, Vince!

9
by Coach B (not verified) :: Fri, 08/12/2011 - 6:20pm

This just sounds random with no statistical support. Where is this random "divide by 20" coming from. And the multipliers, how were they determined. When you start factoring in your own numbers, that is a red flag for bad data or data fitting. I love metrics, but I don't see how this one flies.

10
by DJensen (not verified) :: Thu, 04/12/2012 - 2:41pm

Nice work! A few questions come to mind.

1) What was the original formula that considered an "8" the goal number?
2) Do you have the numbers for this years class of Receivers?
3) Is there a similar formula for Running Backs?
4) What would the ideal number be for Tight Ends?

Thanks.