Evaluating My 2016 Young WR Model Performance

Prior to the 2016 season I created a model that tried to predict PPR production on a per-game basis for young wide receivers, which I defined as WRs in their first four years in the league. This article is going to break down the model results. For all the data the models were built from, as well as the 2016 model projections themselves, refer to the linked article.

Reviewing the Models

First I’m going to start with a brief review of the models. I created a separate model for each of the first four seasons of an NFL receiver’s career. The factors I looked at included:
  1. Year (SEAS)
  2. Logarithm of draft position (L.DPOS)
  3. Draft age z-score (AGE.Z)1
  4. Final year collegiate market share z-score (MS.Z)
  5. PPR points per game the prior year (PPR.N)
  6. Games played the prior year (GMS.N)

Rookie Model

The rookie year model had the following significant predictors:
Term Estimate p-value
Intercept -367.866 0.0045
SEAS 0.189 0.0033
AGE.Z -0.276 0.1706
MS.Z 0.482 0.014
L.DPOS -1.689 <.0001
(SEAS-2010.15)*(AGE.Z+0.402) -0.136 0.0503
(SEAS-2010.15)*(L.DPOS-4.598) -0.115 0.0835
In this model, age was more important than market share (on a scaled basis). This model produced a backtested R-squared value of only 0.298 with a RMSE of 3.357 on 318 observations. Projecting rookies is really hard.

Second Year Model

Here’s the model for second year receivers:
Term Estimate p-value
Intercept 0.765 0.1757
GMS.N 0.260 <.0001
PPR.N 0.461 <.0001
AGE.Z -0.686 0.0045
MS.Z 0.882 0.0002
(GMS.N-9.450)*(PPR.N-6.060) 0.063 <.0001
(AGE.Z+0.533)*(MS.Z-0.406) -0.658 0.0109
This model has a backtested R-squared of 0.514 with a RMSE of 3.451. The term for SEAS dropped off. Also, in a player’s second year, we should give slightly more weight to his college market share than his draft age. Prior year PPR points per game and rookie year playing time are the most significant factors.

Third Year Model

Term Estimate p-value
Intercept 4.818 0.014
GMS.N 0.243 0.012
PPR.N 0.391 <.0001
MS.Z 1.029 0.0001
L.DPOS -0.874 0.0042
(GMS.N-11.348)*(PPR.N-7.596) 0.072 <.0001
This model produced a backtested R-squared of 0.513 and an RMSE of 3.641. Notice the AGE.Z term has dropped off completely, meaning by a receiver’s third year in the NFL we shouldn’t even factor in draft age. Draft position, NFL production, and collegiate production tell the whole story.

Fourth Year Model

Term Estimate p-value
Intercept 0.924 0.3075
GMS.N 0.219 0.0069
PPR.N 0.499 <.0001
(PPR.N-8.611)*(PPR.N-8.611) 0.027 0.0382
The R-squared for this model is 0.467 and the RMSE is 3.748. In this model model, any notion of draft position, age, and college production has dropped off. This is the point in a receiver’s career at which he has fully established (or failed to establish) himself as an NFL receiver.

Model Results

Alright, let’s jump into how the model fared.
  1. Taken from Jon Moore’s Phenom Index  (back)

Subscribe to the best value in fantasy sports

You're all out of free reads for now and subscribing is the only way to make sure you don't ever miss an article.

By RotoDoc | @RotoDoc | Archive

Comments   Add comment