Joining a Dynasty League? Here are 30 of the NFLs Greatest Young Stars.

Pictured (From left to right): Jalen Hurts, Patrick Mahomes, Justin Jefferson

Introduction:

Fantasy football has become a beloved pastime for millions of Americans each year. In recent times, the emergence of “dynasty leagues” has taken the fantasy football experience to new heights. In a dynasty league, participants draft their dream team, which they’ll manage not just for a single season but potentially for the next five, ten, fifteen, or even twenty years. Drafting in a dynasty league carries substantial weight; the choices made can significantly impact a team’s performance. The age of the players drafted is a crucial factor, as younger talents offer the prospect of a longer tenure on your roster.

For some enthusiasts, fantasy football isn’t just a hobby — it’s a passion. They invest countless hours and a substantial amount of money each year in pursuit of the perfect draft. The outcome of their fantasy football season can have a real impact on their lives, influencing whether the lights stay on or whether there’s food on the table. While many enjoy fantasy football for its entertainment value, for some, it’s a way of life, and the draft is no laughing matter.

This article sets out to provide data-driven insights into identifying young players with the potential to follow in the footsteps of today’s fantasy football stars. Last year, the top players in the three primary positions were Patrick Mahomes (QB), Justin Jefferson (WR), and Christian McCaffrey (RB). These players made a game-changing impact for fantasy football owners in 2022. By leveraging statistics from the 2022 season and employing data science techniques, we aim to pinpoint ten promising young players in each main position (QB, RB, WR) who are most likely to ascend to similar stardom. We hope our audience uses these predictions to surpass their competition and become the best in their league with the help of data.

Data

The data used for this analysis came from the Pro Football Reference website using NFL statistics for the season of 2022. It was extracted and uploaded to Kaggle two days ago by user Jalin Gordo with the title ‘Fantasy Football 2022 Season’. This dataset has statistics on players such as their age, team, games played, and all in-game statistics such as yards, touchdowns, and attempts.

Process

To gather an analysis on what players are going to be successful in the future, we decided to first look at what players are currently the top of their class and find what current players (under the age of 25) have the closest cosine similarity to them by using the sklearn python module. This will give us an insight into players that may be on track to eventually rise to the caliber of Patrick Mahomes, Christian McCaffrey, and Justin Jefferson.

The cosine similarity of these players will be conducted based on their in-game stats exclusively. That being said, the ‘Team’, ‘Position’, and ‘Age’* fields were not included in this analysis.

A problem that arose early on was the lack of values in certain fields. This was solved by filling in the missing values with the mean of all players in that position. The section of code used to do this is pictured below.

Figure A — To be referenced below

After the insertion of this section, our program was ready to run.

Results

Patrick Mahomes:

The top 10 players under the age of 25 with the most similar trajectory (cosine similarity) to Patrick Mahomes are as follows:

1. ) Tua Tagovailoa
2. ) Justin Herbert
3. ) Trevor Lawrence
4. ) Mac Jones
5. ) Davis Mills
6. ) Kenny Pickett
7. ) Jalen Hurts
8. ) Zach Wilson
9. ) Brock Purdy
10. ) Justin Fields

Christian McCaffrey

The top 10 players under the age of 25 with the most similar trajectory (cosine similarity) to Christian McCaffrey are as follows:

1. ) D’Andre Swift
2. ) Najee Harris
3. ) Antonio Gibson
4. ) Josh Jacobs
5. ) Rhamondre Stevenson
6. ) AJ Dillon
7. ) Kenneth Walker III
8. ) Travis Etienne
9. ) Dameon Pierce
10. ) Cam Akers

Justin Jefferson

The top 10 players under the age of 25 with the most similar trajectory (cosine similarity) to Christian McCaffrey are as follows:

1. ) CeeDee Lamb
2. ) Amon-Ra St. Brown
3. ) DeVonta Smith
4. ) Ja’Marr Chase
5. ) Jaylen Waddle
6. ) Tee Higgins
7. ) Josh Palmer
8. ) Brandon Aiyuk
9. ) Jerry Jeudy
10. ) Terry McLauren

Advice:

These players are provided based on their similar trajectories to the statistical leaders of their given positions. That being said, please do not use this as a cheat sheet or an ultimate source of fantasy football knowledge. Use this information as a tool to help make the most educated decisions possible when selecting your team.

Limitations

While this list provided players with the most similar cosine similarities to the best of the best, a few things should be noted.

First and foremost, it’s important to note that these similarities do not necessarily stem from factors like a player’s speed, strength, agility, injury history, playmaking abilities, or raw athleticism. Moreover, they do not account for variables such as the coaching staff, supporting players, or the team schedule of the players. Instead, these similarities come from the type of role a player fulfills on their team, rather than the sheer quantity of plays they are involved in.

To illustrate, consider the case of Antonio Gibson, who may not have been ranked among the top 10 running backs in the 2022 season. Nonetheless, he was frequently utilized in Washingtons passing game in a manner similar to how Christian McCaffrey was heavily involved in San Francisco’s passing plays. This emphasis on their respective roles and responsibilities within their teams forms the basis for the observed similarities. That being said, the their cosine similarities are also heavily derived from the players productivity as well.

The second limitation we are aware of is in figure A (pictured above). While it is typically good practice to plug in the mean for a stat that has a blank value, we plugged these in without taking player position into consideration.

For example, runningbacks do not typically pass the ball at all. If the runningback were to have a blank value for passing yards, this should be filled in by using the mean passing stats of all running backs (near 0). However, that is not the case. Using the code in figure A, the runningbacks passing yards stat would be filled in using the mean value of ALL players, giving them stats that are very likely not to be accurate.

The link to the repository used can be found here