smart data for smart decisions
At footballytics we combine football & data analytics skills and have many years of tactical analysis & scouting experience in professional football. Football and data and not data and football.
Use data to make evidence-based better decisions in scouting and match analysis.
We help football clubs integrate, structure, refine and interpret match event data according to their own match philosophy. To maximize the potential of data and turn it into valuable and individual new insights and competitive advantages. The goal of a modern club should be to consistently make good decisions based on the best possible data.
See blog post > Data Analytics in Football
Data Scouting Platform
With our own data scouting platform, we interpret millions of event data and make players comparable in a simple way. The insights are used in scouting and in match and opponent analysis to make better data-based decisions. Noticeable patterns sometimes also allow conclusions to be drawn in relation to the coach..
Data analytics does not aim to replace football scouts. Rather, data analysis provides a valuable complement to the talent identification skills of the scouts.
An additional tool. A good scouting process consists of different phases and starts with data scouting. This is so that you don't narrow down to individual players too early and can really consider the entire player potential at the beginning. In addition, data does not have favorite teams or players.
The risk of perception error is minimized by the "second opinion" of the data. See blog post > Datascouting
Break away from reactive scouting for agent recommendations
In our own scouting solution we access a huge amount of data (Wyscout) with about 150 leagues and over 100'000 player entries. The more than 100 individual metrics (dribblings, shots, submissions, duels, passes, success rates, interceptions, etc.) have been combined, condensed and refined into 16 smart key metrics. This makes it possible to compare the performance data of the players quickly and easily.
Scouting practice: data driven scouting
We show you an example of how our "data driven" scouting solution works. We are looking for the following player profile:
· a young technically skilled offensive winger
· who is excellent in delivering goal chances
· can workout chances for himself
As a supplement he should:
· be implemented good in the team build-up
· often be able to intercept balls
· often win defensive duels
In addition, the player should cost less than 1 million. One million is really little, for a good, creative winger, who can prepare and also score goals. We are looking for the "gold" in football, so to speak.
Data scouting best practice
Now we start with the actual data scouting and begin to feed our platform with values. At the beginning, without filters, we have a total of 68'336 players from the 2020-21 season.
I do not use the search/evaluation according to the number of goals or assists. Because the efficiency in the conclusion can vary strongly after season and an Assist is not only the merit of the assisst giver, but also of the goal scorer. For these statements, Expected Goals (xG) and Expected Assists (xA) are certainly the most meaningful metrics.
Important: In data scouting, we do not want to "downscout" to 5-10 players, but to about 50-100. Because the data is only one dimension in the consideration. Data scouting is only about an initial selection. The selected players are then to be analyzed in detail via video. So that to the player values also the context can be formed and the how and why questions can be answered. So it pays not to define the criteria too high, so that there is room for the important context. But I guarantee that data scouting will uncover numerous surprising players that would otherwise not be on your radar.
Data scouting with our data driven platform
To start, our application has the following default filter: 2020-21 season, all leagues, all positions, no restrictions. You can scout from a total of 68'336 players
Main Position Offensive Wing OR Attacking Midfield First we define the position in the search. We take wingers and offensive midfielders. Filter: Attacking Midfield OR Offensive Wing Result: 15'092 players still match the criteria
Elimination of statistical data outliers To eliminate the statistical outliers we set the minutes played to > 500 Experience shows that performance stabilizes after 500-800 minutes of play. Result: 8'272 players still meet the criteria.
Young perspective player We are looking for a talented player with perspective and development possibilities. Therefore we limit the age to 19 to 24 years. Result: 3'555 players still meet these criteria.
Now the actual selection begins with the player skills.
The most important scouting criterion is my coach
First of all, I consider which player skills my current coach can develop further and which skills are difficult to improve. The coach also plays an important role for us in scouting. If he is able to improve the players and develop innovative solutions, I can buy much more cheaply on the market and later earn more from the sales. This is the desired business model of the vast majority of clubs. The coach also plays an important role in terms of business management.
Dribbling / Dribbles I start with dribbling as the first priority. This skill is extremely important on all positions defensively we offensively. However, it is very difficult to improve after a certain age. In our platform we have indexed all the key values of the players per position on a scale from 0 (underground) to 100 (world class). With the index values, players are easier to compare.
I set the dribbling value to at least 45, which is already a more than good value. The winger is over 45 compared to all other wing players. I also have in mind that the player must not cost more than 1 million in the end.
So he certainly can't pass like Verratti, dribble like Traoré, cross like Kostic, lay up like Messi, and finish like Lewandowski. Not yet 😊 Result: 846 players still meet these criteria
Good scouting also means recognizing the harbingers of the future
Benchmark: Best top5 league dribblers are: Traoré (Wolverhampton), Neymar (PSG), Cherki (Lyon), Ontiveros (Huesca). .
Delivery Now we take care of the skills of the templates. This is another skill that cannot be improved extremely. Since you have to find creative solutions and make quick decisions during the game. I set the Delivery index value to at least 35.
The players we are looking for are perspective players. So it is logical (also for the sake of the market price) that skills are not fully developed. Result: 403 players still meet the criteria
Benchmark: Best assist providers Top5 leagues: Illicic (Atalanta), Kostic (Frankfurt), Di Maria (Real Madrid), Golovin (Monaco).
Shots The completion/shots index does not mean how good the shots are, but how often and how promising the player is to finish the goal. The ability to get into position to score is a skill that can be greatly improved with good and fresh ideas. If I have a good coach who encourages innovative runs and collective interactions, I can tend to use a lower value here for scouting. A good coach will also bring central midfielders and outside backs into the fold. I set the Shots Index value to at least 40 Result: 159 players still meet the criteria
Benchmark: Most frequent goal finishes wing top5 leagues: Neymar (PSG), Jota (Liverpool), Berardi (Sassuolo), Salah (Liverpool).
Efficiency in finishing I'll leave out the efficiency in finishing for the moment. Because this value can vary strongly depending on the phase. And it is not my first priority for the player I am looking for.
So, now we have found 159 players, who are good dribblers and good assistants and can also score goals themselves. Not bad.
Complementary skills Reminder. As a complement, he should:
· offer himself well in build up
· often be able to intercept balls
· often win defensive duels
Build up / Passing I set the Pass Index value to at least 25 . This guarantees that the player is involved in the game, gets into position well and has numerous ball contacts. Result: 142 players still meet the criteria
Benchmark Winger with the most passes Top5 leagues: Rafinha (PSG), Cairney (Fulham), Fabregas (Monaco), Coutinho (Bercelona).
Interceptions I set the interceptions index value to at least 25. This guarantees a respectable number of ball recoveries. By the way, we show the number of interceptions "possession adjusted". I.e. the number is proportional to the possession. So that players from teams with little possession are not disproportionately rated. Result: 130 players still meet these criteria
Benchmark: Wings with the most interceptions Top5 leagues: Leckie (Hertha), Haidara (RB Leipzig) , Musiala (Bayern Munich), J. Victor (Wolfsburg).
For fine tuning, I'll set some more values to filter out the worst ones in the corresponding categories.
Defensive Duels I set the Defensive Duels to at least 35. This sorts out the players weakest in two fights. It wasn't a criteria to begin with, but I refine my list with it. Result: 93 players still meet these criteria
Benchmark: Wings with most defensive duels won Top5 leagues: Rafinha (PSG), Del Castillo (Races), Djénépo (Southampton),
Market value - from dream to reality The results now include players like Coman, Chiesa, Cunha, Diatta and Hauge who cost 10 to 65 million. So I set the Market Value parameter to a maximum of 1 million.
Final Data -pre scouting result TOP 63 At the end of our data driven scouting process, there are still a whole 63 young perspective players with a lot of great skills in the selection. We started with 68,336 players and a full 99.91% of the players did not "pass" the selection criteria.
Each and every one of these remaining 63 players is certainly worth a closer look.
Data driven scouting is pre-scouting. In the following video scouting, it is also important to question the values considering the main and secondary positions. There is always the possibility that the good values were collected on an alternative position and thus can distort the result. Example: A winger plays as an alternative position in the defensive midfield. He collects good tackling values, which are credited to the main position, because the data does not allow anything else.
It is noticeable that not a single player from the top 5 leagues is among them. This is not a surprise, as such players from the top5 leagues cost much more than 1 million.
In the detailed analysis, of course, the strength of the league must also be taken into account, because the league strength can already influence the values by 5-20%.
The following players from Switzerland made it into the TOP63: Liridon Balaj (Aarau), Alexis Antunes (Servette, needs a review because he only played around 600 minutes due to injury) and Sayfallah Ltaief (Winterthur).
From Austria represented are: Atsushi Zaizen, Lukas Fridikas (both Wacker Innsbruck. 2. League). Among others also Philip Otele (Kauno Žalgiris) appears, who is traded at the moment with FC St. Gallen and FC Zurich. The best dribblers of our selection are: Aron Dönnum (Vålerenga Fotball), Marlos Moreno (Lommel SK), Philip Otele (FK Kauno Zalgiris), Davo Fernández (Atlético de Madrid B).
The best in delivery are: Shakhboz Umarov (BATE Borisov), Zymer Bytyqi (Konyaspor), Matan Hozez (Maccabi Tel Aviv) and Nico Williams (Bilbao Athletic B)
The best at creating goalchances are: Deabeas Owusu-Sekyere (Paide Linnameeskond), Philip Otele (FK Kauno Zalgiris), Nadrey Dago (FC Sheriff Tiraspol) and kas Fridrikas (FC Wacker Innsbruck).
S. Umarov (Bate) Z. Bytyqi (Konyaspor), P. Otele (Kauno Zalgiris), Davo Fernandez (Atletico Madrid B) and Ferreira (Grémio) are our top 5 selected players.
Here is our final scouting list with the TOP 63 along our offensive index (highest values on top), AM = Attacking Midfield, OW = Offensive Wing (main position)
Top 6 comparison
Conclusion This scouting process took us 20 minutes and some brain power using our own scouting solution. It was kept generic for simplicity. Of course, in a real case, we will go into more detail. Still, it is quite impressive how easy it is to draw insights from the data once it is aggregated and refined.
If clubs have individual position-specific needs, we are able to adapt the interpretation logic quickly and flexibly.
After the data scouting (pre.scouting) we would analyze players via video sequences in detail along our offensive index. This would result in a shortlist of about 20 players. After that, whole games are to be analyzed so that the performance can be assessed in detail and also character & mentality.
We would not be surprised if some of the players in our selection change clubs next season.
Blog from www.footballytics.ch About data analytics topics in football - smart data for smart decisions
share this post