Home » The dream of every football manager: an algorithm to never make mistakes

The dream of every football manager: an algorithm to never make mistakes

by admin

There is an invisible side of football revolutionized by technology, a way of reading games profoundly influenced by data collected on the pitch and processed by powerful algorithms.

A computer can, in real time, give effective indications to the coaches, such as to condition their choices beyond intuition and tactical competence? Not only is it possible but these are the notifications that, for example, someone sitting on the Sassuolo bench could receive during a match:

! 20’ pt. The willingness to receive the pass by Berardi on the right wing is increasing

  • “Berardi has not yet received the ball in that area”

! 30’ pt. The pressure applied by Sassuolo on the opponents is increasing by 20% compared to the first quarter of an hour

  • “The pressure increases, Sassuolo is growing”

VAR, Goal Line Technology, augmented reality: in recent years the game of football has undergone an extraordinary technological evolution. The new tools available make it possible to collect an ever-increasing amount of data.

It therefore becomes of fundamental importance to transform “big” data, that is of large quantity, into “smart” data, that is to say few and useful indications extracted through processing and analysis. This is the goal of the Virtual Coach project, a tool developed by Math & Sport, a spin-off of the Politecnico di Milano.

Starting from last season, Lega Serie A has given the possibility to all clubs to use the Virtual Coach. Through a system of cameras in the stadium, the positional data of the players and the ball are collected, as spatial coordinates on the football field, about 20 times per second. Mathematical models and artificial intelligence algorithms process this raw data to build physical and tactical indicators in real time who support the technical staff in analysis and decisions.

See also  Covid: brain alterations discovered one year after the infection

In modern football, in the classrooms of Coverciano as on TV, we frequently talk about the attacking styles of a team, comparing those that offer a proactive kick – through maneuvering – with those that wait for the right moment to quickly restart on the counterattack, in an attempt. to surprise the opponent.

To quantify and measure this different approach has been very complicated up to now, due to the enormous amount of situations and actions in a game. In the 380 Serie A games played this season, there have been almost 60,000 actions, of which over 50,000 maneuvered and over 8,000 in transition.

Thanks to the Virtual Coach algorithms, this type of analysis is now possible and available to technical staff and coaches (even in real time). The system manages to categorize every single action at the very moment in which it is taking place. And it saves all the key information, such as the average speed of the players, the duration and effectiveness of the maneuver.

Thanks to this innovative system it is possible to study the game in a deeper and more detailed way. In fact, very interesting correlations and trends emerge, such as those proposed in the following table:

RANKING MANEUVERED SHARES

Rank

Team

Average number Maneuvered

1

JUVENTUS

77.7

2

ATALANTA

77.6

3

LAZIO

76.4

4

SPICE

76.1

5

FIORENTINA

75.3

6

SASSUOLO

74.8

7

ROMA

72.9

8

NAPLES

72.2

9

CROTONE

71.6

10

BOLOGNA

69.6

11

INTER

69.5

12

GENOA

69

13

MILAN

68.9

14

TORINO

68.7

15

VERONA

68.5

16

PARMA

67.9

17

SAMPDORIA

67.7

18

UDINESE

67.5

19

CAGLIARI

67

20

BENEVENTO

58

See also  How often to fertilize orchids in summer? Here's how to do it right

The top teams in terms of the number of actions handled are Juventus, Atalanta and Lazio. Compared to the positions occupied in the final classification of Serie A 2020-2021, the “surprises” are Spezia and Fiorentina respectively in fourth and fifth place in the special list compiled by the Virtual Coach, and Crotone ninth. In reality, these teams prefer the consolidation of ball possession over quick restarts. At the bottom of the list there are teams that prefer to verticalize more with long balls, therefore opting for a lower number of maneuvered actions than those involving transitions, as can be seen thanks to the following table:

RANKING OF RESTART SHARES – average number of transition actions per game

Rank

Team

Average No. of Transitions

1

LAZIO

13.3

2

CROTONE

13.2

3

SAMPDORIA

12.9

4

TORINO

12.5

5

UDINESE

12.5

6

BOLOGNA

12.0

7

PARMA

11.9

8

GENOA

11.9

9

ATALANTA

11.9

10

ROMA

11.8

11

VERONA

11.5

12

MILAN

11.1

13

FIORENTINA

10.9

14

BENEVENTO

10.8

15

CAGLIARI

10.5

16

SPICE

10.4

17

INTER

10.1

18

NAPLES

10.0

19

JUVENTUS

9.9

20

SASSUOLO

9.2

The first teams of this special classification, such as Sampdoria, Udinese and Turin, are generally voted to immediate verticalization once they have regained the ball. On the other hand, the teams that do not prefer restarting actions are generally those that occupy a higher ranking position in Serie A (Inter-Napoli-Juventus) or typically devoted to consolidating possession (Sassuolo).

DURATION OF ACTIONS vs TEAM SPEED (average player speed)

In this graph all the maneuvered actions have been considered

It is noted that in general the high-ranking teams have a longer duration of the actions, combined with a lower team rhythm. This may be an indication that these teams prefer a more orderly and less hectic maneuver than other teams in the league. In particular Sassuolo is the team that performs very long maneuvered actions on average (+ 23s). The exception is represented by Atalanta, which unlike the traditional big players in the league, does shorter actions on average, but with a high team pace.

MANEUVERED ACTIONS: DANGER vs DURATION

In this chart, only the most dangerous actions of each team have been considered

Danger is a proprietary index of Math & Sport which takes into consideration the position on the field of the players and the ball carrier to understand the potential danger that can generate that specific action, regardless of the outcome of the goal.

There is a clear separation between the first-tier teams, which are the most dangerous and on average make prolonged actions, and the second-tier ones, which produce shorter and on average less dangerous actions.

Within the two groups, however, there are differences: while Atalanta, Milan and Lazio prefer shorter actions (under 22 seconds on average), Rome, Naples, Juventus, Inter and Sassuolo prefer longer actions. It is interesting to note that in these special rankings it is Sassuolo (almost 26 seconds on average) who build actions with a longer duration.

The relationship between duration and dangerousness of the actions somewhat belies the common thought according to which Inter would have won the 2020-2021 title thanks to the immediate verticalizations on Lukaku. The maneuvered actions of the Nerazzurri last, according to the algorithms of the Virtual Coach, on average well over 20 seconds, a sign that the team has a complex game construction with many preparatory steps for the decisive play.

The other interesting aspect that finally emerges is that Atalanta and Sassuolo have had, most of all in the last Serie A championship, a very characterized style of play. Gasperini, Atalanta coach, having based the effectiveness of his team on high aggression, does not need prolonged actions to reach great danger.

The Bergamo team has in fact an average duration per action in line with that of small clubs, but with respect to these its actions are often developed entirely in the third offensive, that is to say in the field area manned by the opposing defense, therefore with a rate of much higher incisiveness.

On the contrary, Sassuolo has built its identity of the game, and its strength, on the continuous possession of the ball, starting from its own penalty area, deliberately lowering the pace of the game and arriving calmly and with all the effects in the offensive half. Anesthetizing the opponent and then hitting him at the right time was the formula adopted by De Zerbi and which transpires from the data collected.

.

You may also like

Leave a Comment

This site uses Akismet to reduce spam. Learn how your comment data is processed.

This website uses cookies to improve your experience. We'll assume you're ok with this, but you can opt-out if you wish. Accept Read More

Privacy & Cookies Policy