The analysis from this research provides a number of interesting results. We have combined an overview and understanding of the transfer market and its economics, politics, and case studies with data analysis of transfer market trends over time and modeling the relationship between game performance and transfer fee.

From the overview of the transfer market, we find that the transfer market is a huge part of the game of soccer, for players, managers, owners, and fans alike. The amount of money decided on for a transfer fee involves personal relationships, agents, location, club funds, and other political factors. The transfer market, unlike major American sports, had experienced an explosion of cash with no limit over the last few decades. Economic factors such as merchandise sales, advertising, ticket sales, and fan appeal play a huge role in acquiring and selling players.

The trends of the transfer market are clear: the amount of money spent has exploded exponentially from 1992 to 2019. In the past, the Premier League was about split on the number of clubs that made a profit from transfers and the number of clubs that lost money. Now, almost all clubs spend more than they bring in. Additionally, the most money is paid for middle aged players in their prime, with younger rising talent slightly more expensive than older, retiring stars.

The first major question we tackle is what game statistics explain transfer fees. In other words, which game performance outcomes have the biggest impact on transfer fees. Using multiple different methods, we determine that goals and assists per 90 minutes have the biggest influence. This suggests that clubs may pay a premium for star forwards and attacking midfielders. A star defender or keeper would produce just as good results for the club at a lower price.

The second major question is whether we can model the relationship between the important game statistics and transfer fee. Although we had challenges with limited data, we create a machine learning regression model using a neural network, that was able to rather accurately predict Eden Hazard’s transfer fee of 90 million Euros. Its tendency to underestimate the other transfers, however, indicates that the previously discussed economic and political factors are common and play a significant role in increasing transfer fees. Thus, a manger or club looking for the best value may look to under the radar players that don’t come with the extra costs. Ultimately to win championships, the club would want to buy players whose market rate closely matches the rate indicated by their game performances.

Finally, the case studies on major transfers illustrate the combination of all these findings in real world examples.

For more information on this topic, check out the related work page.