The Model
The goal of this model is to use college football statistics to predict fantasy football success for a college running back prospect in the NFL. The variable we are predicting is total fantasy points (PPR) within an RBs first three seasons. If a player is still active in the NFL and has not reached three years then their current historical points have been extrapolated to three years.
The Results
The Adjusted R-squared for the model is 0.6216. Each metric is represented as a career seasonal average. The most predictive metrics are:
- Weight from 247
- Receptions
- My homegrown Rush Yards Over Expectation (RYOE) metric
- EPA Per Reception
- Break Away Run Rate (% of carries for >=10 yards)
Initially, I did not have weight within my model but including this has improved the Adjusted R Squared over 0.5, which means over half the variation in the model can be explained by the metrics used as inputs. The preferred weight for an NFL RB is 210 lbs. Based on some research in the Twitter thread below we can expect an RB to gain between 10-15 lbs while in college, which means an RB who is at least 195 lbs entering college should arrive at the optimal weight. An RB with extreme production can break through the normal weight barrier such as Darrell Henderson, Kenneth Gainwell, or Deuce Vaughn.
What I could deduct of this:
1) The average weight prospects gain is 10-15, more is a strange case. Stop projecting a 40 pound gain (this is a Devone Achane Tweet).
2) If you are 180 or less as a HS prospect, chances you make it to the NFL day 1/2 are really slim (8.3%).— Adeiko_FF (@Adeiko_FF) April 28, 2021
We are attempting to predict fantasy points in a PPR scoring format so the ability to be a good receiver is weighted heavily. Volume is more predictive than efficiency. Receptions are represented as receiving volume and EPA per reception is represented as receiving efficiency. Having both gives a low-volume receiver who was extremely efficient, like Derrick Henry or Nick Chubb, an outlet to earn a top spot on the model.
Rush Yards Over Expectation is the third most predictive metric. Expected rush yards are based on yards to goal and win probability. In general, a player who has a starting point further from the goal (larger runway) and with a higher team win probability will have a higher expectation for rush yards. Win probability is an important input to expected rush yards because it provides a real-time indicator of how much better a team is than its opponent. Using the strength of schedule can be a lagging indicator for comparing opponents within a certain game. Also, using yards to goal is an important input because we can assume a defensive scheme and defenders in the box based on the relation to the goal line. Below I have provided the ten highest expected yard situations as well as the ten lowest expected yard situations.
Finally, I have adjusted the break-away run rate metric from the percentage of runs greater than or equal to ten yards, which was previously fifteen yards. The game is played in smaller chunks than 15 yards and has proven to be more predictive within this model.
The Prediction
I have provided three charts below. The top 50 current and prior prospects, the top college prospects, and the top 2022 prospects, which has been arranged by predicted fantasy points. I would let draft day create a tier of prospects for upcoming rookie drafts and organize the tiers based on the projection. For devy/C2C drafts I will be using ADP data to create tiers and organize these players based on the projections. A few notes from the results:
- Canton Notes:
- Travis Etienne has Christian McCaffrey’s upside if healthy.
- Kevin Harris is the sleeper prospect of the 2022 class. He draws parallels to James Robinson as a prospect. I would like to see Harris on LAC alongside Austin Ekeler.
- Max Borghi and Kyren Williams have been pushed to the top by extremely high reception totals. I do like the trajectory of Tyler Badie over these two prospects in upcoming rookie drafts.
- Campus Notes:
- TreVeyon Henderson is projected to be a better prospect than Ezekiel Elliot.
- Lew Nichols has the chance to be the prospect we had anticipated for Jay Ajayi. He most likely will miss Day 2 of the NFL draft but could be a solid NFL prospect.
- Trevion Cooley is the same prospect as Raheim Sanders, take the 86 pick discount and select Trevion Cooley instead.
- Kendre Miller is my top target in 2022 Devy Drafts. His projection is minimized by his lack of receiving ability but he is a former high school QB who had a prior top RB prospect on his team. With Zach Evans moving on he should get the full opportunity to make or break his career as an NFL RB through improvements as a receiver.
