2024 NFL Draft Analysis

I have spent countless hours making, optimizing, and producing predictive models for practically every single position in this year's class. I gave a success metric value, which I called a LUBIN Score, chance of becoming a star, and chance of becoming a bust for over 200 prospects in this year's class through these advanced predictive model.

Before showing all of the results of this year's class, I wanted to prove the accuracy of my models, so I made a plot for it to be easily visualized:

The predicted success scores and actual success scores have a correlation of .685

It can be seen from the plot that there is a strong correlation in my model's predicted success and actual success. The model is not perfect, as "all models are wrong, but some are useful" - George E. P. Box, and I believe my model is one of those useful ones.

There are a couple final notes before getting into the specific results of the models:

• The only position that does not have a model is quarterback, as I spent hours trying to create one, but never got a successful enough result to publish it.

• Some positions take into account combine results, so if a player did not compete in the combine, then they might not be present in the results.

• I did not have sufficient data for any player who did not play in Division 1 college football.

• In order, the positions I have most faith in are: LB, TE, Safety, DL, Edge, WR, RB, CB, and OL.

Now to the fun part. Here is a graphic of every single drafted player with their LUBIN Score, Star Probability, and Bust Probability:

If you are interested in reading about any specific player or any specific position, I posted separate articles that give thorough analysis on each position's results. These articles also have each position listed from highest LUBIN Score to lowest LUBIN Score, which could help to relativize.

Using the graphic above, we can see how good each team did in the draft by their individual selections . Going team by team, most teams drafted someone who has at least a decent chance of becoming a star, giving hope to fans of practically every team. Also evident is that many players have rather high likelihoods of becoming a bust, which should be expected, as most players in a 7 round draft do not make it into great players, but I guess one could hope.

Some teams had a wide range of drafted players, with some projected to be stars and some projected to be busts. What I wanted to see now was which teams were best as a whole:

Reminder that no QBs were involved in these scores

This graphic shows that the Raiders had the best draft in terms of pure LUBIN scores. They also had the highest probability of their players becoming stars, as well as the lowest percentage of their players becoming busts. I included the average pick spot in the graphic because the lower number your pick spot is, the easier it is to get better prospects and subsequently a higher LUBIN score because you would have better draft capital. Taking pick spot into account, the Broncos and the Dolphins had really impressive drafts with LUBIN scores over 56 and an average pick spot over 139.

In reality, most players drafted after the first few rounds will not pan out to having great careers, and that is reflected in the scores. Because of this, I wanted to see how well teams did in the first two days of the draft, which includes only rounds 1 through 3:

Reminder that no QBs were involved in these scores

Looking at this graphic, the Tennessee Titans blew all other teams out of the water in terms of their early round draft picks. They drafted JC Latham and T'Vondre Sweat, which the model believes to have an average LUBIN Score of 88.8, which is outrageously high. The Broncos, which were the 6th best drafting team according to the all-rounds graph, place dead last in the graphic above of the first three rounds. Given that the model does not incorporate QBs, their selection of Bo Nix in the first round is not taken into account. The Packers and Commanders are interesting as each they drafted 5 players, meaning they had a lot of opportunity to draft great players, but also a lot of opportunity to mess it up. Looking at the model's results, it seems as if both teams did a solid job with their selections, with both teams having an average probability of players becoming stars above 20%.

These models took me north of countless hours to create and optimize. This was done with the pure intention and hefty goal of predicting NFL success from collegiate athletes with an extreme accuracy. After completing this whole process, I believe that I have completed my goal.

Feel free to reach out with any questions about my methodology, specific player evaluations, etc. My Twitter/X account is @JonahAnalytics.