Analytics & Datascouting: Setup, Opportunities & Challenges

"Having access to people who can understand data is critical. It's the most underrated feature in football." Arsene Wenger

At last year's StatsBomb conference at Stamford Bridge in October, various data analytics experts gave exciting and practical presentations on topics in sports analytics. One of the pioneers and head of analytics at Liverpool FC Ian Graham gave a presentation on how to maximize the impact of analytics at a club.

Graham's first experience in football was back in 2007 at Tottenham Hotspur with his first data analytics projects. Since April 2012, the has headed the analytics department at Liverpool FC.


"Data analytics is the process from data to decisions. Data is reviewed, cleaned and interpreted - with the goal of gaining valuable information and insights to make better decisions."

Jürgen Klopp showed just how wonderful a tool analytics is with Liverpool FC. He was open to new things and combined his emotional intelligence-based leadership style with the capabilities of Ian Graham's analytics team.

Meanwhile, everyone doesn't want to miss the boat. Early adopters such as Liverpool, Midtjylland, AZ Alkmaar and Brentford have gained competitive advantages through years of experience and learning. Meanwhile, there is probably almost no ambitious club where data scouting and data analytics is not an issue.

The main message of Graham's talk was to focus as much as possible on analytics work that makes a difference and creates value. That may sound like common sense, but at the same time it's also a bit difficult to grasp and also difficult to implement.


"Basically, it's about giving limited resources to data analytics and insights as much power as possible."

Impact is achieved by using analytics simultaneously in different areas such as scouting, game analysis, athletics, medicine, and also in marketing and designing products and the customer experience.

But this can also be achieved by involving the team, empowering people, asking the right and important questions, addressing the right problems and setting the right priorities. And then finally, implementing the measures consistently and efficiently and reflecting on their success.

"Because we often don't have a clear strategy, we collect and analyze mountains of useless data, hoping to miss as little as possible."

Accordingly, it is clear to us that we need to connect data analytics with strategy.

All priorities and every single measure should be thought top-down from the strategy of the association. To develop power and directness. And this turns the Analytics starting question for the associations completely upside down. We don't start with the data, we start with the strategy.


"At the beginning, don't ask yourself what data can benefit you. Ask yourself what data you need to measure and manage your strategy."

Data Analytics is like possession in football. It is not an end in itself and it is not for beauty. Every action, every pass. every action should have a purpose. With the goal of strengthening the implementation of the game strategy piece by piece.



"Scouting is 95% rejecting players and 50% preventing bad transfers".



Scouting is 95% rejecting players

This thesis is easy to explain. Football is business. Many are earning through the whole food chain. Every semi-professional club gets several sensational players offered by agents and consultants every day. For top clubs, this can be as many as 15 per day. Just processing and responding to these recommendations takes a lot of time (and nerves).

The scouts in the clubs do not have the time to deal with all these recommended players in detail. Let alone watch them all on video or even in the stadium. Nor should they, as reactive scouting on agent recommendation is probably not a sensible strategy.

Clubs depend on a qualitative and quick initial fast assessment of the offered player. Hopefully, as many clubs as possible are already doing this in a resource-saving and qualitative way with a single data check.



A. Knauff footballytics data profile
A. Knauff footballytics data profile

After a quick look at the right data, it quickly becomes clear whether the player fits the profile and whether it makes sense to analyze him further in detail. See our blog Data Scouting.


Due to the sheer amount of daily recommendations, only about 5% of all offered players survive the first data check and are further analyzed by the scouts.




Good scouting is only 50% successful

The first thought on this is "What - only 50%?" Surely a professional scout should aspire to be right in almost all cases.

But it is like in business, that the success of a transfer depends on many different also complex factors. And not all of them can be correctly assessed in advance with a high degree of probability.

In his presentation, Ian Graham used some basic calculations to show how difficult the selection of players can be. Using a simplified example, he explained that the ratio between a successful and unsuccessful transfer is only about 50:50, despite professional scouting and good assessments. Even if you have a 90% certainty for a group of categories.

Even if the scout does a super job and his key assessments are 90% reliable, the success rate of a transfer is only 50%.

Is the player really that good? Does the player fit into the playing system? Does the coach like him? Is he really fit and that fast? How strong is he mentally? How does he react in crises? Can he assert himself? Is he a team player? Is he capable of learning? Can he continue to develop?


See following explanatory graph for Ian Graham's statement (image by The Athletic).


Graham's scouting (image by The Athletic)
Graham's scouting (image by The Athletic)


It is amazing (but not for the math) that when we are 90% correct in our assessment of six dimensions, the transfer is only 53% successful. Every wrong transfer is a big mortgage for the club. And if the scout does not do a good job, the success rate will be even lower.




Data analytics challenges for football clubs

Every introduction of a new technology brings exciting possibilities and great opportunities. But also some challenges, which we would like to discuss in more detail.

As always, the challenges lie not only in technology and engineering, but also in people and culture.


"Progress and success comes less from solving problems and much more from taking opportunities." John Naisbitt




Intelligent investments in software AND know-how

As always, money plays an important role. In many cases, clubs have very limited resources available for new topics.


"It is not enough for a club to just buy good data and analytics software."

Every club should consider how much money it invests in software and how much money it invests in experts and know-how building. Nowadays, people are fascinated by data and data analytics and everyone wants to take advantage of the new opportunities. They all want to work with the best software and the best products. But they underinvest in experts who understand the new business, the new technology and khow how to make the data talk.

The tool is important, but the expert does the artwork.

Once you want to leverage the use of data in more than one area, you need deep technology, IT, security and analytics knowhow.


"An efficiency rule of thumb is to invest 80% in experts and 20% in software to get the most out of it."

But not all clubs, like Liverpool and Manchester City, have the ability to hire entire departments of Data Analysts and Data Scientists.

Data Analytics can be used very broadly in different areas such as scouting, match analysis, athletics, medicine and also in marketing and shaping the customer experience.

We recommend that clubs create a new and dedicated Analytics role within the club and staff it with an expert. This person is the interface for all club areas on the topic of data usage. Don't think that these skills can be acquired along the way. It takes years to become a good data analyst. And please don't delegate analytics to someone who is already overloaded with a thousand other tasks.

If you can't fill a new position for financial reasons, then get an expert to help you in the initial stages so that a good setup can be chosen.




Anchoring in strategy and empowerment in the association

Smart and efficient data use cannot function in isolation and detached from the other areas. The use of data is not an end in itself. To develop clarity and power it needs a connection to strategy. Every company and every football club has a strategy. And in the future, this will additionally need a data strategy.



Culture eats data for breakfast Data analytics doesn't want to replace anyone or take anything away from anyone. It's an additional great tool that complements the skills of scouts and game analysts.



It's not just about using data to help make decisions. It's about empowering other people to deal with data. It's about establishing a culture that values data and systematically uses it as a decision-making tool, right down to the very last corner of the club (including at the very top). It is also about dealing openly with fears. Dealing with criticism and resistance.



"An honest and open approach and view of one's own system and culture is the foundation for continuous development.


We also see that data use in the clubs often begins in scouting, detached from the other areas. However, this is not the case all the way to the top. Ideally, the systematic use of data is advocated, launched and supported by management right from the start. This increases the impact and helps immensely in difficult moments.

"In many cases, new methods and technologies fail not because of the technology, but because of the culture and the lack of commitment in management."


What Data Analytics can achieve depends not only on the technology, but also on how it is embedded. But also on the embedding in a larger context and culture.

We recommend that organizations make Analytics an integral part of their strategic and operational business. Already there where the strategy is defined.




Data scouting - every data set has its own reality

For data driven scouting, you first need the data. Meanwhile, all clubs have access to the data. Compared to the amount of data, there are also very cost-effective solutions available. The choice of provider is mainly based on the data availability of specific leagues. These tools provide a strong data foundation that scouts can use for analysis.

Some solutions also offer intelligent algorithms that aggregate and evaluate player scores. However, it should be recognized that each data set, each collection methodology, and yes, each rating algorithm has its own logic, imprint, and truth. This truth varies from data provider to data provider.

"We recommend that clubs use data from at least two different providers for data driven scouting."

This multi-provider approach is a bit more involved but not much more expensive. But it minimizes risks, makes the data statements more robust, and you are less dependent on the quality of a single provider.


And a contradiction of the data view to the subjective perception is already an added value. Since you are motivated to look closer at the data and the player.

On NASA manned missions, important instruments are duplicated. And these must have been designed and built with different technology and logic for safety reasons.



state of the art - combination of event and tracking data

With the event data you "only" record the players and the actions on the ball. The remaining players are not considered. If you consider that a player has the ball for only 2-3 minutes during a game and 21 of 22 player are not considered - you come to the conclusion that 95% of the game is not included.



With the combination of event and tracking data it is possible to capture all players.

This is done by capturing the game with a camera and building a 2D model of the game with software.




Then we have the data/movements of all players. If you now add the velocity of the GPS data, you also have the velocities. The output is a 2D model in which all players move and all ball and noBall actions can be analyzed visually and with data.


The result is the largest possible DataAnalytics data foundation.


Eventdata + Trackingdata + Velocitydata = 100% of game and 100% of possibilities.


This takes DataAnalytics to another level. Now, running paths without the ball, distances between players and chains, attacking patterm or even coverage shadows and pitch control , body alignment can be analyzed.


With the combination of event and tracking data, in addition to the ball actions, the remaining 95% of the game can be captured.

The disadvantage lies in the very high investment required by hardware, software, DataScience and operations resources, which seriously implemented from over half a million. Club-specific custom metrics can also be developed.


Data itself is not valuable to people. Its the ability to use data who generates insights about things the people care about.


The question is whether a top club that invests millions monthly in player wages should not invest 1% of that to make the players even better or to be able to sign even better players.




Future - Scouting with Artificial Intelligence (AI, AI)

There is a relatively new category of scouting tools on the market that advertise Machine Learning and even Artificial Intelligence (AI, AI). Powered by AI are the taglines.

For example, this tools are saying to be able to predict a player's market value. In simple terms, the AI makes predictions based on the data, checks over time whether the prediction is accurate or not. Depending on the result, it independently adjusts the logic and learns from it to improve the predictions.


These AI tools are still in their infancy. And where AI is on it, AI is not always in it. The quality of these solutions is very difficult for an association to judge. We have not yet been able to form a detailed picture of whether the promised potential is realistic.


In any case, human skills that cannot be automated will become increasingly valuable in the digital future. In any case, we are a long way from algorithms deciding which system to play, which player to sign, who to substitute and when. And we should not strive to get there either. At the center is and remains the human being with his know-how and experience.


Can you keep your cool despite public opinion and make your own decisions with conviction, regardless of external pressure? If you can, then #DataAnalytics is for you. If you can't, then why bother? (Oliver Seitz)


"80% of clubs scout 20% of the same players. With soccer knowhow and data driven scouting & DataAnalytics, you can break the pattern and find strong undervalued players that the others don't have on their radar."

And then one day when all clubs are scouting with data and analytics?


Then you should be scouting the best data analysts :-)



Smart #DataAnalytics is self-financing and saves money! If you invest a few percent of players' wages, you will be able to hire much cheaper & better players.

Did you like our blog? To never miss a post you can subscribe to the blog here. Or share this post on social media via the buttons at the bottom.





footballytics - we know how to make the data talk

We combine the competences of football tactics, scouting and data analytics and support clubs in interpreting and using data to make better decisions in scouting and match analysis based on data validation.




 

Blog von www.footballytics.ch About Data Analytics in football - improve the game - change the ǝɯɐƃ

Share this Post

0 Kommentare

Aktuelle Beiträge

Alle ansehen