Predict Player Free Agency Market Value

Business Problem

Each offseason, contracts signed by pro sports free agents direct the flows for hundreds of millions of dollars and alter the fates of billion-dollar sports franchises. These deals are often the result of competitive bidding contests in which teams try to accurately value the unknowable, highly volatile future value of the player and outbid their competitors without overpaying and wasting money.

The risk of under-bidding and not valuing the player highly enough is the lost production that the bidding team will miss out on from that player’s contributions on the field, as well as the competitive advantage that a rival might gain from acquiring that player’s services. If the player truly is as good as the hype, then losing out on their talents can set the team back for multiple seasons and put them at a competitive disadvantage. Thus, bidding enough to realistically win the player’s services is important.

However, the risks of over-bidding are just as imposing, if not more so. There are countless stories of massive contracts signed by players who never lived up to their market value and some who barely even saw the field in their new uniforms due to injury or ineffectiveness. In these cases, the teams committed massive sums with little-to-nothing to show for it; resulting in embarrassment, scorn from fans, and little ability to acquire other stars while their budgets were consumed by these high-dollar flops.

Thus, the savvy front office must be able to 1) understand the player’s future value as well as possible, and 2) understand the marketplace for that player’s services to develop effective off-season strategies and bid intelligently for the players that will create the most return on investment.

Similarly, players and their agents must develop realistic expectations for how the market will play out to build their own negotiating strategy. This will help them set reasonable opening asks, recognize the best offer when they see it, and generate demand in the market for their services to maximize the final contract value.

Intelligent Solution

An informed view of how the free agency market will play out before it even starts can help both the teams and the players to construct their strategies. There are enough players, teams, and historical seasons to build ML-driven models that are accurate player production forecasts and can identify patterns in market values for players. By combining these two predictions, players can get a better understanding of their market value and teams can optimize their bidding strategies to create a winning group on the field of play.

Forecasting Player Fair Value

Using historical performance data, models can be built to make the best-possible prediction on future player performance. Depending on the sport, there can be massive amounts of data and quantifications of player characteristics to inform these projections. Once player value is determined, teams can set up market comparisons for similar players and determine fair value for each player in the free agent market instantly and without engaging an entire team of analysts to build these valuations one-by-one.

Similarly, players can use these projections to build their case for their asking price, pointing to the value they will bring their future teams and going-rates in the marketplace for their level of talent.

Forecasting Player Market Value

The exercise above for forecasting player value is good for determining fair value; but market value is oftentimes irrational and deviates from reason. ML models can take fair value and future player production into account, but also can consider market demand and supply for similar players to determine if it is a buyer’s or seller’s market before bidding ever begins. Additionally, past ‘irrational’ behaviors by teams can also be factored in to create a more prescient image of who will sign where and for how much. This insight can help players evaluate offers and help teams set bidding strategies to avoid overpaying in blind auction scenarios.