Finding the Best Undervalued MLS Contracts in 2024

In this article, I delve into the model I developed to assess player contracts and highlight the five most undervalued deals currently in MLS.

How do I determine whether a contract is undervalued?

Getting the Data

A big shoutout here to American Soccer Analysis. Their interactive tables offer clean, accurate player data with advanced metrics and salary information dating back to 2013 (they also have excellent explainers on their site for many of these metrics). This served as the foundation for building my model.

Preprocessing the Data

Before building the model, I needed to preprocess the data, with a focus on the unique salary structure of MLS. To evaluate players in 2024, I excluded this year from the training and test sets. Next, I filtered out goalkeepers, as their role differs significantly from outfield players.

The next major adjustment involved removing Designated Players (DPs). This MLS rule allows teams to sign two or three marquee players whose salaries exceed the cap without impacting it. For example, Inter Miami can pay Messi a $12M base salary despite the league's $5.7M salary cap. Since DP salaries can vary dramatically based on a team's market size and spending power, and that volatility isn't the focus of this analysis, I excluded these players.

Finally, I accounted for changes in the salary cap over time. For instance, the cap in 2013 was around $4M compared to $5.7M today. To address this, I calculated each player’s salary as a percentage of the cap. This adjustment allowed me to use the entire dataset while maintaining a consistent measure of a player's value relative to their team.

Building a Model

Building a model to tackle this question was not easy, particularly given the limited size of the training set, which contained just over 3,100 observations. I experimented with three different models: LASSO Regression, Ridge Regression, and Random Forest. Thanks to its flexibility, Random Forest outperformed the others in handling this complex problem.

The Random Forest model outperformed LASSO and Ridge Regression, with a lower Mean Squared Error (MSE) and a higher R² score. As a result, we chose Random Forest as the model to evaluate player valuations.

Feature Importances

When it comes to on-field performance, minutes played stands out as the most significant feature in predicting a player's salary.

Who are the players that stood out based on this model?

There are many effective ways to interpret this model's results, but one of the most compelling insights was identifying undervalued players—those whose predicted salaries, based on on-field performance, exceed their actual earnings. For 2024, the top five standout players include Andrew Privett (Charlotte FC), Jackson Ragen (Seattle Sounders), Luca Orellano (FC Cincinnati), Max Arfsten (Columbus Crew), and Lucas Bartlett (DC United). Interestingly, four of these five players were drafted through or connected to the MLS SuperDraft, highlighting its potential as a valuable platform for uncovering underappreciated talent.

Note: This is not a comprehensive scouting report on these players. There are gaps in the data that don’t capture every aspect of a player’s game. Instead, this is a general overview, highlighting some strengths and weaknesses for each player mentioned.

For each player, there is pizza plot presenting KPIs that illustrate their 2024 season as percentile ranks compared to other players in their cluster.

Andrew Privett - CB (Charlotte FC)

Privett’s base salary for the 2024 season was $71,000, but his expected salary—based on performance metrics—is estimated to be approximately $424,000. This indicates that Charlotte managed to save about 4.3% of their salary cap. While I will delve into the numbers behind Privett's on-field performance later, I will first examine his minutes played, as this is a significant factor influencing the model's outcomes. This season, Privett logged an impressive 3,044 minutes. For context, players in his salary range typically average 930 minutes—a staggering 226% increase! To compare Privett’s on-field performance to that of similar players, I utilized KMeans Clustering to segment the dataset. The analysis determined that eight was the optimal number of clusters. Unsurprisingly, Privett’s cluster primarily consisted of center-backs, aligning with his role on the field.

The pizza plot highlights that Privett has performed at an above-average level compared to other players in his cluster this season, showcasing the exceptional value he has provided to Charlotte given the value of his contract. The data reveals few significant weaknesses in Privett’s game. However, there is room for improvement in two areas: delivering more penetrating passes and improving the accuracy of his more challenging attempts.

Jackson Ragen - CB (Seattle Sounders)

Ragen’s base salary for the 2024 season was $98,000, while his expected salary—based on performance metrics—is estimated at approximately $425,000. This indicates that Seattle managed to save roughly 4% of their salary cap. Looking at Ragen’s minutes played this season, the Sounders CB logged an impressive 3,244 minutes. For comparison, players in his salary range typically average 1,250 minutes—a 160% increase! Similar to Privett, the players in Ragen’s cluster were primarily center-backs, reflecting his position on the field.

Ragen’s performance in the top ranked Seattle Sounders defense (in terms of goals against) is slightly less balanced compared to Privett, particularly in the metric *Interrupting Goals Added*. This statistic, developed by American Soccer Analysis, evaluates the value of a player’s defensive actions, focusing primarily on on-ball contributions. Notably, it does not account for key defensive aspects like positioning and marking.
While Ragen excels in playing out from the back, the data highlights areas for improvement in some of his defensive responsibilities, such as clearing the ball and winning tackles.

Luca Orellano - WB (FC Cincinnati)

Orellano’s base salary for the 2024 season was $90,000, while his expected salary—based on performance metrics—is estimated to be around $406,000. This means FCC managed to save approximately 3.85% of their salary cap. Despite joining mid-season as a replacement for Álvaro Barreal, Orellano logged an impressive 2,905 minutes. For context, players in his salary range typically average 1,250 minutes—a 127% increase! Unlike the center-back comparisons above, Orellano’s cluster comprised of wingers and wingbacks, aligning with his role on the field.

Orellano impressed in his debut season in MLS, showcasing his ability to generate goalscoring opportunities. He frequently acquired the ball in the attacking third and converted those opportunities into goals through his strong shooting and dribbling. However, his metrics for receiving the ball didn’t stand out—a reasonable outcome given the comparison to traditional out-and-out wingers in his cluster.
Orellano’s passing and playmaking metrics were roughly average, so this is an area he could focus on improving in the upcoming season. With the uncertainty surrounding star Lucho Acosta’s future, if Orellano is able to make strides in these facets of his game, this will really shore up some of the creative side of FCC and provide service to MLS’s largest incoming transfer ever in Kévin Denkey.

Max Arfsten - WB (Columbus Crew)

Arfsten’s base salary for the 2024 season was $71,000, while his expected salary—based on performance metrics—is estimated at approximately $370,000. This means that the Crew saved roughly 3.6% of their salary cap. This season, Arfsten took a major leap within the first team for the Crew, logging an impressive 2,184 minutes. For context, players in his salary range typically average 930 minutes—a 134% increase! Similar to Orellano, Arfsten’s cluster consisted of wingers and wing-backs.

In his first full season with the first team of the Crew, Arfsten showed he can be productive player within this league. He posted above-average dribble numbers, despite playing in a more supporting role to the stars of the team in Cucho and Diego Rossi. Look for Arfsten to take a big next step this season, especially if a big club like Flamengo or Monterrey offer Columbus a fee they can’t refuse for Cucho.

Lucas Bartlett (CB) - D.C. United

While Bartlett received a base salary of $98,000, his expected salary is estimated to be around $390,000, meaning that D.C. have saved roughly 4% of their salary cap. He recorded 3150 minutes played this season and players around his salary range typically receive 1248 minutes, a 152% increase! Like Privett and Ragen, the players that made up Bartlett’s cluster are center backs.

From the data, it appears that Bartlett is an average center back in MLS. He doesn’t do a whole that’s well above average, but doesn’t commit a lot of fouls and can occasionally complete a difficult pass. He has some areas where he needs to improve though, specifically with his on-ball defensive duties such as clearing the ball and winning tackles. Interestingly, in a D.C. United team that has Christian Benteke, by-far the best aerial duel threat in MLS, Bartlett opts to play short and avoid long vertical passes.

Honorable Mentions:

Patrick Agyemang, Moise Bombito, Duncan McGuire, Ian Murphy, and Andrés Gomez.

Conclusion

Thank you all for reading! Going through this whole process of selecting features, building a model, and producing visualizations to explain the results has been a very useful and practical project. I am hoping to continue to dive into this dataset that I’ve created and have a couple articles lined up that I am excited to get to work on! If you are interested in my work, check out my GitHub where I’ve shared various projects, models, and visualizations. Also, feel free to connect with me on LinkedIn or Twitter—I’d love to chat. See you next time!