Forecasting 2024 NFL Rookie RB Success: 3-Year Model

May 22, 2024

Kirby Lee-USA TODAY Sports

The 2024 NFL Draft is now in the books, giving us dozens of new offensive players to consider for our fantasy teams. In this article, I will discuss some of the value picks available at running back in rookie drafts.

Our friends at Dynasty League Football (DLF) have already compiled some post-draft rookie ADP, and Jonathan Brooks is the unanimous top running back off the board. All four models I use rank Brooks as the top running back in this draft. While the timeline of his ACL recovery makes his unclear for early this season, his long-term prospects for dynasty managers look excellent.

More NFL Rookie Forecasting: WR | LB

Forecasting Running Back Success

As in the past five seasons, I will estimate the odds of each player putting up a top-24 season within the first three years of their career by using a combination of analytical models. While the models we have used in past seasons have certainly found some hidden gems, they have not managed to match the success of the receiver model, which seems to find an undervalued late option almost every year. For that reason, I put in some time this offseason trying to further improve the RB models in hopes of catching up to the WR models.

The most striking thing about the core WR model is that it is incredibly simple, using just two variables to score each receiver: draft pick and career market share of team yards from scrimmage. Despite that simplicity, it has continued to perform exceptionally well. With that in mind, I went searching for a simple and accurate predictor of RB success. If you want to read the full details of what I tried, I described the process on my personal blog. The short version is that the most accurate model was again one using just two variables: draft pick and age. That model outperformed thousands of other models I tried, models that considered hundreds of different RB metrics.

In addition to the simple model, I am also including three others. The first is a model that projects using different criteria depending on each player's categorization into the passing, big, or lighter running back "cluster". I described that model in more detail in the 2019 RB article. The last two models use more complex machine learning approaches (SVMs and random forests). Despite their sophistication, neither model outperforms the two previously mentioned, so they receive a lower weight in the combined scores below.