Is Splashing Cash on Transfers Really Worth it in the EPL?

An data analysis project looking at the affects of player transfers on the results of the top EPL teams across seasons

What is the English Premier League and who are the best teams?

The English Premier League is made up of the top 20 English club soccer teams. It is one of the oldest leagues in the world and comprises some of the best teams and players in the world. It is quite competitive as the EPL generally has the smallest spread between the top half of teams across any league in Europe. Therefore, teams needs to improve and add to their squads every year to stay competitive. Over the decades, 6 teams have distanced themselves generally as consistently being at the top and these are the 6 teams I will use in my analysis: Arsenal, Chelsea, Liverpool, Manchester United, Manchester City, Tottenham.

My data goes all the way back to 2005–2006 season to the 2017–2018 season. You can see that the top 6 teams have significantly more wins than the rest as I have kept Everton (7th) for comparison. You can also see that these same top 6 teams have the 6/7 greatest number of incoming players.

Here you can see that teams need at least 20 wins a season to be competitive for a trophy in the top flight. It also seems that the average number of transfers per year for these teams seems to be about 5. Now we will take a deeper into each team’s wins and their transfers

When should teams buy players and how long before they make an impact?

There are two windows where teams can buy players in Europe. The main transfer window is in the summer and the second smaller window is the middle of season around the holiday time. However, since each team’s style and league style is so different, a player that was bagging 2 goals a game in France might struggle to find that same form in England and vice versa. Therefore, we will look at cycle of transfers and how and when they affect wins.

At first glance there may not seem to be a pattern, however, looking at these graphs more closely, there seems to be spikes and dips in both lines that seem cyclical. If we take Arsenal (top left) as an example, you can see that the spikes in the number of transfers (pink line) precede the spikes in the number of wins (blue line). The pink like spikes about two years before the blue line. This type of patter can be seen at different frequencies across the teams. These different frequencies can reflect the type is takes players to adapt to the team’s system play and get their feet under them. Therefore, generally teams should invest in key player at least one season before they are going to need them because there seems to be a lag in the results after a player is bought.

How can you maximize the value of the players you buy?

Many clubs in recent years have been dishing out tens of millions of dollars are players from all over Europe. However, we see that just because you buy the most valuable players doesn’t mean you will be successful. Furthermore, just like how the Oakland A’s in baseball tried to find value by buying undervalued players or how investors try to value invest lets look to see if and how some players are either overvalued or undervalued. I first build a linear regression model to see the affects of age and the different positions below.

There is definitely some difference between each position with defenders and goalkeepers on the cheaper end while wingers and forwards are a little more expensive. It doesn’t seem super significant though. Now lets looks to see if players are overvalued or undervalued. I am going to compared their transaction fees to their estimated market value (what other teams would pay) to see if there is some opportunity to get ahead. If the transaction fee is fee is greater than market value then they are overvalued and vice versa. Below is the summary of my logistic regression model to predict which players are overvalued and which players are undervalued.

First, you can notice at the bottom of the image the comparison between my model accuracy of (.53) compared to putting in all 0s (.42). Yes, my predictions aren’t super accurate but they are more than the baseline. Furthermore, we can see how age has a very significant affect on if a player is overvalued or undervalued. There is a negative coefficient for age which means that as the player gets older they are less and less overvalued. This makes sense because players just like other assets are valued on their future potential and growth. Therefore, you will be paying a steep premium for a player that is young and full of potential but maybe hasn't proven anything. Young players also have a good resale value because they still have much longer to play unlike older players who may retire soon and are more injury prone. These components are reflected in the price of players at different ages.

Secondly, you will notice that certain positions are significantly overvalued than others. Each factor here is being compared to the Attacking Midfielder who is slightly overvalued already. You see that Central Defenders, Goalkeepers, and Center Forwards are greatly overvalued. This is interesting because we see that defenders and goalkeepers are priced lower than the rest but compared the the market they transact at a much higher rate. These three seem to make sense because they can be directly attributed to statistics. In soccer, there very few tangible statistics used to judge the skill of a player. It is very subjective as each players plays different and fits a team’s system and style differently. Therefore, it seems to make sense that these three category of players has objective criteria (goals against for defenders/goalkeepers and goals scored for forwards). Teams that are in the market for these players are willing to pay more because they can show that they are will either allow less goals against them or score more goals for the teams. That direct impact seems to be valued more.

Key takeaways

Overall, teams want to buy players before they need them because it takes some time for them to get used to the team and leagues style and finally start to help deliver wins. Transfer do help though because without transfers you won’t be competing for any trophies. However, you don’t have to splash cash on young players with lots of hype because they are usually the most overvalued. If you are going to spend money on young players you should by players that aren’t playing Center Back, Forward, or Goalkeeper because they will be the least overvalued. Teams should look to convert similar positioned players into these positions to save money on buying overvalued players on the market. Transfer help you replenish your leaky bucket of talent and let you compete at the next level if you have the fund but may need to be more tactical with player and position selection if you have limited funds.

Data

[embed]Football Results, Statistics & Soccer Betting Odds Data Historical Football Results and Betting Odds Data 25 seasons results | 18 seasons betting odds | 18 seasons match stats…www.football-data.co.uk

[embed]Top 250 Football transfers from 2000 to 2018 List of 250 most expensive transfers that took place between 2000 and 2018www.kaggle.com