Football IQ Observatory


Statistics as a major driver for talent identification & scouting purposes: The Football IQ approach

There’s been a lot of talk lately about how data-analytics can be used to enhance identification and scouting capabilities. Having worked consistently for more than 3 years on building Football IQ Index, perfecting our algorithms, pioneering the field of harvesting big-data for talent identification, scouting and recruitment, we know exactly what it takes to build a comprehensive and objective analysis framework.

– Identification of top-performers
– Monitoring of player performance during the season and/or across different seasons & consequent tracking of improvement/decline
– Identification of players with similar performance “fingerprint”

are becoming the norm, in a market that has been slowly but steadily shifting to the new “big-data” paradigm.


Can statistics be used for scouting? They sure can, but what if we are talking about players used in different positions? Then, the context becomes entirely different.

And what happens with the team’s playing style? Obviously, a player who plays for a team that applies a ball-playing style of play, will rank higher in most passing metrics than a player who plays in a team which predominantly counterattacks.

Additionally, there is the playing time parameter. For prospective key signings teams usually prefer signing players who have featured in more than 75% of their previous team’s matches, whereas for future prospects they don’t always pay so much attention on how much they’ve played, rather on their performance when they were on the field. Considering the variation in requirements, flexibility cannot be treated as an “if” but as a significant prerequisite.

Last, but not least, lets suppose we want to compare players from different competitions. For this to be done efficiently, competition dynamics must be taken into consideration.

Another important aspect has to do with “reading” the data. Usually, the metrics used are on a per 90′ format. So there, we might have a statistic that a player attempts 3.4 dribbles per 90′. How can you incorporate
a) position dynamics
b) team dynamics
c) competition dynamics
d) playing time 

into that 3.4 dribbles per 90′ metric?

It’s not easy. And even if someone was able to do it, the specific format is not user-friendly at all and you can’t go far with it.

Common Scale (1-100)

Applying the transformation to the 1-100 scale, we create data-sets that are built to act as dynamic comparative assessment indicators. This way, when we see a player with rating of 74 in e.g. dribbles involvement (attempted) we know immediately that the player attempts more dribbles than 74% of all players in the same competition, used in the same position. Moreover, that 74 rating already embodies the principles of team dynamics.


What’s also extremely important is a further contextualization which offers significant analytical advantages:

– Involvement (attempted)

– Productivity (successful attempts)

– Efficiency (success percentage)

Suppose we want to examine what is the playing style of a given player i.e. what does he try more often. For this we’d need to check out Involvement metrics. If we want to quickly check the areas where a player not only makes a lot of attempts but is also very productive, we’d check Productivity. 

Productivity though, can sometimes be misleading, that’s why its also very important to always check Efficiency metrics. Efficiency metrics can also be used as points of reference: a player attempts very few dribbles per 90′ but his efficiency is very high, maybe there is a “window of opportunity”.


It’s always good to be able to go into the fullest detail possible. However, sometimes -for different reasons- we might want to check just one number. For that, we are producing the overalls. In simple terms, overalls are weighted indicators combining involvement, productivity & efficiency in a single rating, emphasizing on the ratio between involvement & efficiency. Sounds simple, right?

Summary KPIs

The most “potent” of our performance indicators, incorporating overalls across all categories (passing, attacking, defending). It is directly associated with the level of versatility of each player and it is widely used to sort players who have exhibited high levels of performance in more than one areas.

Summary KPIs (Leo Messi 2018-2019)

Domestic Performance Indicators (DPIs)

We give our clients the freedom to select the data-set that best suits their objective at a given time, by producing two different versions of Domestic Performance Indicators:
-Simple DPIs
-Weighted DPIs (taking into consideration playing time)

Universal Performance Indicators (UPIs)

For incorporation of competition dynamics, we produce the Universal Performance Indicators (UPIs). In simple terms, after taking into consideration objective facts about each competition (financial status, performance in international competitions etc) and more importantly performance data related to the predominant performance quality and style of play of each competition, our algorithm adjusts the weighted DPIs, providing a much more balanced data-set.

Suppose we are looking at the DPIs of e.g. the Greek Superleague. In the DPIs of each competition, there are lots of players with ratings of 100, 99, 98 and so on.

When it comes to UPIs these ratings get adjusted to the level of the Greek Superleague compared to the other competitions around the world. So, a 99 DPI rating achieved in the Greek Superleague would become a 80-82 Universal Performance Indicator. This approach allows for a much cleaner and straight-forward ranking of player performance on a global scale.

Similar players & matching algorithms

The common 1-100 scale was chosen for another reason: It works great with our “similar players” algorithm, while at the same time rendering the “matching player to requirements” process a walk in the park, allowing us to maximize suitability.

A potent road-map

Applying such an integrated approach transforms statistics into a multi-dimensional road-map, much like what a sonar does in the depths of the ocean, identifying reefs and icebergs, along with their depth and space they occupy.


Can statistics alone, even through the scope of an integrated analysis framework such as Football IQ Index, tell us everything about the player? Not really, as important factors such as the mentality of a player, are not taken into consideration.

They can definitelly be used as guides for fast and objective identification and performance evaluation, but the player must be assessed further. That is why technical scouting plays a pivotal role for further assessment of player ability and behavior.

We’ll go into more detail about our technical scouting principles here at Football IQ, and how our analysts operate for the creation of complete player folder in next week’s post.

Thank you for reading and stay tuned!

If you found this article interesting feel free to like, share and/or comment!

Share on facebook
Share on google
Share on twitter
Share on linkedin

Leave a Comment

Your email address will not be published. Required fields are marked *