The most complete midfielder - Would you have expected him to be number one?
Photo by Tom Fisk from Pexels

The most complete midfielder - Would you have expected him to be number one?

Or: What happens if you would ask data instead of the yellow press.

I recently stumbled upon an article that tried to rank the world’s best midfielders in football. It was said that midfielders play various roles on the field, and are probably the most important players in terms of building up attacks. Further, they are supposed to have an all-round view of the match, and as most of the action takes place around their area of the pitch, midfielders often exert the greatest degree of control over the style a match is played. Indeed, midfielders, by definition, act as the junction between offense and defense. At this point, pundits and other self-professed experts often talk about “the complete midfielder”, who masters the art of offense and defense equally well. While skimming over the list of players, I was certainly able to easily agree that all these guys belong to the best what football has to offer. But then, I started to ask myself why would I agree that all these players are “complete midfielders”? What makes a midfielder a complete one? And, who am I to agree with such a list without even knowing who of them scored the most goals, provided the most assists or won the most tackles?!

And, who am I to agree with such a list without even knowing who of them scored the most goals, provided the most assists or won the most tackles?!

Finally, I began to think about this in more detail. If I was an analyst, how would I approach the question of identifying the most complete midfielder? At first, I decided to narrow the sample to my domestic competition, the German 1. Bundesliga, which I certainly followed most thoroughly this season. With a continuously growing fascination, I investigated a couple of publicly available data sets. There are a bunch of figures, potentially helpful to distill the 1. Bundesliga’s most complete midfielder of this year’s season (2019/2020). After running some tests, I decided to isolate a few meaningful metrics and boiled it down to a number of five attributes that appeared to define a complete midfielder in both offensive and defensive dimensions:

  • xG 
  • xA
  • Key Passes
  • Tackles Won
  • Interceptions

In order to better compare the players among each other, I came up with the idea to calculate just a single score out of these five core metrics. To preserve a high level of measurement, the value of each player in each dimension was compared to the dimension’s average among the peer group of 1. Bundesliga midfielders with more than 30% of total minutes played. By doing so, you will get a number for each dimension that states how much a player is better or worse than the average. Summing up the differences then estimates the completeness of the respective midfielder and assigns a single performance score. For those - like me - who are preferring the visual representation here’s the formula:

formula performance score
variables formula performance score

That way, we standardize the average at 0, simply meaning that the average player achieved a value of 0 in each dimension, resulting in a total score of - no kidding - 0. Please note, due to the construction of the formula, players who perform worse than the average receive negative values. Using the sum of all dimensions avoids to “overvalue” for example a very offensive midfielder, creating an enormous xG value but not winning any tackle, so this guy does not automatically outstrip his more defensive peers as he will probably receive negative values for the rather defensive-related dimensions. This should virtually balance the different roles midfielders can have on the pitch. Another way to take account for different roles and positions would have been to adjust the samples, so comparing offensive midfielders separate from defensive ones - but as we are looking for the most complete guy, I thought that was not the most appropriate option. Judging from the amount of time you already invested in reading all this, I bet you’re curious who actually turned out to be the most complete midfielder of this year’s 1. Bundesliga campaign (based on data until week 21; sample: 91 midfielders). Here you go:

Third place, with an overall score of 447:

performance chart player 3

Second place, with an overall score of 652.69:

performance chart player 2

The most complete midfielder of 1. Bundesliga so far in 2019/2020, with an overall score of 657.67:

performance chart player 1

I know it’s mean but the final disclosure will follow later on (colors used give no indication whatsoever) ??…

While browsing through the final table of results another question emerged in my brain: How did those midfielders perform relative to their market value and if there was a hidden gem player? In order to answer this, I enriched the data set with market values and calculated the difference between the performance rank and the market value rank. Among the top 20 players of the performance ranking have been three surprising guys who would have been expected in far lower areas of that table - according to their market values. Those guys are:

table surprise players

Please find below the scatter plot of the final result set in two different representations. The first graph is showing market values and performance scores on the respective axis, while the second one uses the ratio of the performance score and the market value.

scatter plot 1
scatter plot 2

Please let me know if you are interested in one specific result. Moreover, I would be super curious to get to know your expected top 3 players from this little ranking exercise.

Here is the “solution” ??:

table top players

Of course, the conducted analysis has its weaknesses. Nowadays, it becomes more and more complex to even distinguish between second strikers and attacking midfielders, wingers and right or left midfielders or even offensively oriented right/left-backs. For sure, there is also a more precise way to define the dimensions to look at, or in general to tell what it is about to be a complete midfielder. Like the two previous articles, this one is made for the sake of fun and the joy of digging into football data.

I hope you enjoyed reading as much as I did exploring the data and writing down my thoughts.

Let me know if you have any feedback, critics or just which players came to your mind first. 

Thank you!

__________________________

Market values taken from transfermarkt.de

Matthias Werner

I turn disconnected processes into data-backed growth engines | RevOps | BizOps | Data & Analytics

5 年

Sascha Zühlke, ging nun sogar noch schneller als erwartet!

要查看或添加评论,请登录

Matthias Werner的更多文章

  • How I tried to enter the football industry

    How I tried to enter the football industry

    How I tried to enter the football industry 12 Learnings. This is really the most personal piece I ever shared on social…

    5 条评论
  • How to Organise a Football Club

    How to Organise a Football Club

    What are the first steps of a football manager after taking the lead in a new club? I often ask myself, what would I do…

    10 条评论
  • Free Data Scouting Tool

    Free Data Scouting Tool

    How to build your own self-service football intelligence Data scouting is more and more becoming an integral part of…

    2 条评论
  • Leveraging the Academy - The Road to success?

    Leveraging the Academy - The Road to success?

    How younger teams perform compared to the more experienced squads The omnipresent COVID pandemic has a huge impact on…

  • How to tweak your squad through smart budget allocation

    How to tweak your squad through smart budget allocation

    What football can learn from economic theory Reading time: ~15 minutes An efficient resource allocation is crucial for…

    2 条评论
  • Knowledge Acquisition in Football

    Knowledge Acquisition in Football

    Information and knowledge are among the core production factors for modern organisations across industries and…

    2 条评论
  • How to make smarter decisions in football clubs

    How to make smarter decisions in football clubs

    Utilising next-generation data management to increase your club’s return on information Within my last article (Startup…

    6 条评论
  • Startup FC - The opposite of what we have always done

    Startup FC - The opposite of what we have always done

    Nowadays, entrepreneurs and startup businesses are one of the main drivers of innovation in our modern societies. Such…

    8 条评论
  • Gladbach - the real role model in German Football?

    Gladbach - the real role model in German Football?

    Although I am not a fan or emotionally involved, the development of Borussia M?nchengladbach receives my greatest…

  • Why Fortuna Düsseldorf crashed II

    Why Fortuna Düsseldorf crashed II

    Or: Would have been Onisiwo and Mehmedi the better choice? Within the scope of the previous article, it was tried to…

    2 条评论

社区洞察

其他会员也浏览了