The Inferential Analysis of Monthly In-Game Form in Football

The research below includes a complete inferential statistical analysis regarding the monthly form. The sample is consisted of the 12 unique champions in four major European leagues the last 8 seasons. As a key testing factor, a unique weighted statistical ratio has been generated.

Abstract

The goal of this research is to create a measure for the form of the soccer teams using the months as a statistical class (examination periods). The sample includes the last 8 seasons of the leading four championships in the Europe. To measure the form, a statistical ratio has been generated.

The first part of the research compares the champion form scores per each league the last 8 seasons.

The second part contains two sub-parts: At first, the researcher aggregates the league tables and compares them while in the second sub-part the researcher concatenates the individual league tables and compares the ratios for all teams within the sample. The goal of this part is to find the best and the worst performance ratios per each month.

In the final part, the researcher uses the best rated team per month, finds the season that this ratio was the highest and seeks for the MVP. The main goal of the final part is to identify the most important position in the field in order to reach such a high-performance score.

Introduction

Form is a critical element for every potential champion. There are many ways to measure and identify form in soccer. Most of the coaching staffs around the world estimate the performance form using the months as the examination period. During this paper, the researcher will try to compare the monthly form ratings of the 12 latest champions in four major European leagues the last 8 seasons. The goal is to identify the best score per month that the teams should have as a target to measure their form in the field. The scientific method that has been used is the inferential Statistics (Statistics: How to, 2015).

Methodology and Data Collection

The sample for this research includes the required match details for each champion the last 8 seasons (2009-2010 to 2016-2017) in the following leagues: English Premier League, Italian Serie A, Spanish Primera Division and German Bundesliga. The required match details were: goals for, goals against, shots for, shots against. In order to measure the form a weighted statistical ratio has been calculated. The formula to calculate this ratio is the following:

Ratio = 0.6 * (Goals for) + 0.4 * (Shots For) – 0.6*(Goals Against) – 0.4 * (Shots Against)

Example: Dortmund FC is included in the sample only for the seasons that the team has secured the Bundesliga title (2010/2011 and 2011/2012). The final ratio for the team is the average of these two seasons for each month.

There are specific limitations to the research:
1. In order the sample to be comparable the matches that have been played during June and August have been excluded for the sample.
2. Matches in the domestic cups and European competitions have been excluded from the sample.
3. If a champion has more than one titles within this period, the team’s scores have been aggregated using the mean.

The researcher collected the data from online sources, available at the bibliography sector at the end of the research.

Part 1: Team Comparisons for Each League

Part 1.1: Premier League

At this part, the main goal is to identify the best and the worst ratios within each league that is included in the sample. It is very important to examine the ratio scores for each individual league before we expand the comparison to the whole sample in the next parts of the research.

The graph above illustrates the performance scores for each champion in the Premier League the last 8 seasons for each month. There were 4 different champions during this period: Chelsea, Manchester United, Manchester City and Leicester. The results for each month produce some very interesting insights.

Manchester City is the team that is more consistent and always above average. The team of Manchester owns the highest score in the league with a score of 7.92 in May and it is ranks first in the league 7 out of 9 months. Chelsea FC is the second most consistent team. The “Blues” are always above the average except the April, when their performance drop due to lack of motivation after securing a large gap from the other teams in the table.

Manchester United is the disappointment of this sample. The team ranks always below average and only on April the team’s score can be compared with the two teams above. Leicester FC has only one appearance in the sample the season 2015/16. Even though the team won the title, its performance was by far the worst the last 8 years. An interesting fact is that 3 out of 9 months the score of “Foxes” was below zero.

Premier League comparison extracting some very interesting insights. By far the most successful team was Manchester City, while Manchester United was the disappointment and Leicester City FC the negative outlier in the sample.

Part 1.2: Primera Division

In the graph above the examiner can observe the calculated ratios for each title winner for the Spanish Primera Division the last 8 seasons. There were 3 league winners during these seasons: Barcelona, Real Madrid and Atletico Madrid. Some interesting insights are presented below:
Barcelona FC can be characterized as the most consistent team in the Spanish sample because the team follows the average trend-line even though only 1 out of 9 months the team’s score is less that the average. Barcelona holds the best performance score in Spain with 5.35 (December). Real Madrid seems to follow specific form patterns, with 5 tops and 4 lows. Atletico Madrid has produced the worst ratio scores in the sample. The team during the 2013/2014 season has ranked at the bottom in 6 out 9 testing months.

For the Primera Division the results are not so clear concerning the performance winner. Even though Barcelona FC holds the best rating in the league the last 8 seasons, the team of Real Madrid secured 5 out 9 first places while the differences in the team’s performance between the moths were the smoothest in the sample.

Part 1.3: Serie A

Italian Serie A had also 3 unique title winners the last 8 seasons, Inter, Milan and Juventus. The graph above illustrates the patterns for the three winners within the testing period of 9 months. Some critical insights are the following:

Juventus FC has won the Italian championship 6 out of 8 seasons. Therefore, the team’s scores are significantly more accurate that the others. The Bianconeri are ranked first at 6 out of 9 months. Yet, the highest score belongs to Inter the only season that the team secured the title the last 8 seasons (2009/2010) with 6.4 (May). Inter produced also the most abrupt changes over the months. On the contrary, AC Milan seem to follow the trend-line of the average within the sample. The team ranked 3 out of 9 times below the average.

Serie A championship produces less statistical accuracy due to the dominance of Juventus the last 6 seasons. It is truly remarkable though that even though AC Milan had only one presence in the sample, its rating was following the average trend-line on every month. The Italian winner is certainly Juventus due to the team’s consistency in high scores.

Part 1.4: Bundesliga

Bundesliga champions had produced the lowest competition the last 8 years, with only two winners Borussia Dortmund and Bayern Munich. The plot above presents the comparison within the sample. Some statistical hints are the following:

Even though Bayern Munich has secured the title 3 times more than the Dortmund the last 8 seasons, the results regarding the performance are not on-sided. The Munich’s team has secured the first place in 6 out of 9 testing months, having also the highest domestic score with 5.55 (December). Borussia Dortmund participate in the sample in two seasons. In both two seasons the first three months are significantly more inconsistent that the final three. The team presented the lowest domestic score with a calculated ratio of 1.6 (December).

German Bundesliga is also a one-sided championship due to the dominance of Bayern Munich. The two participating teams are aligned only in three months (January, February and May) and unlike the rest 6 months. The best team in Germany is certainly the Bayern Munich.

Part 2: The Inferential Comparisons

Part 2.1: The league comparison

In the first part the observer studies the results for each league in the sample. The researcher has aggregated the results using the mean function to present a final metric (score) for each champion.

The plot illustrates the final scores for each league the last 8 seasons. The statistical insights here are uncorrelated with the title winners, so the results are more accurate. Some of these insights are the following:

Premier League score is significantly lower than the other championships in 6 out of 9 months. The English champions cannot compete with the other champions. Technically, this means that the matches were less one-sided, a fact that in other leagues are less frequent. Italian Serie A champions seem to perform better the winter months while the rest of the season their performance is significantly lower. The Spanish champions have performed above average 7 out of 9 months, with a consistent increasing pattern. The most remarkable score belongs to Bundesliga champions. The German teams, have outperformed the other three leagues 7 out of 9 months. The low score that the teams have on January proves that the winter break ruins the form of the teams due to lack of matches.

The league comparison produces some very interesting insights. The three out of four leagues appear either to be less competitive or produce more clear chances in each match-day regardless the teams. The Premier League can be described as the exactly opposite description with more competition and less attacking production.

Part 2.2: The league comparison

This part includes the big comparison stage for each team the last 8 seasons in the four listed championships. The graph below includes every score for every champion that allows the researcher to find the best and the worst team scores in the sample.

After the examination of each individual league and the comparison of the leagues themselves, the graph above includes the complete view of the sample. There are some very helpful statistical insights in the graph:

Two out of 12 teams are out of competition with the worst scores: Inter and Leicester. The Inter’s score in April and May could only be considered as a statistical outlier. The teams that follow the average European trend-line are the following: Juventus, Chelsea and Borussia Dortmund. The teams with the highest scores are the Manchester City, the real Madrid and Bayern Munich.
The highest score in the sample belongs to Manchester City with a ratio of 7.92 in May followed by a score of 6.4 in the same month (Inter) and the ratio of 6.05 (April – Inter). The lowest ratio ranks belong to Leicester FC (-1.68 – December, -1.60 – February) followed by Inter (-1.36 – February). The winning teams per months are the following:

September – Bayern Munich
October – Bayern Munich
November – Borussia Dortmund
December – Bayern Munich
January – Barcelona
February – Real Madrid
March – Borussia Dortmund
April – Inter
May – Manchester City

Part 2.3: The Optimal Form Pattern

The generated plot below illustrates the optimal ratio score per month that every team should follow in order to increase its chance to secure the title in its championship. This optimal ratio is based exclusively on this research’s sample. The technique of this optimal line is based on Inferential Statistics (Statistics: How to, 2015).

Part 3: The most critical field position for a high score

This part is less quantitative than the rest. More specifically, for each champion-winner per month, the researcher has drilled down into the data, found the maximum ratio for the month with the highest score, extracted the season than the exact month and searched for the Most Valuable Player (MVP) to seek for the most critical field position for each month.
The Seasons that produced the highest ratio scores for each team-winner in part II.I is the following:

September: Bayern Munich – 2015/2016
MVP: Robert Lewandowski
The German striker scored the season with 8 goals in one month (2.7 goals per match)

October: Bayern Munich – 2014/2015
MVP: Xabi Alonso
The Spanish Midfielder scored 1 goal while had 4 assists and 89% passing accuracy and 75% tackling accuracy.

November: Borussia Dortmund – 2010/2011
MVP: Shinji Kagawa
The Central Attacking Midfielder from Japan scored 4 goals, had 23 shots on target and provided 2 assists during the three matches on that month.

December: Bayern Munich – 2014/2015
Manuel Neuer
The German talented goalkeeper had only one conceded goal, while the opponent teams had shooted the ball 16 times on Bayern’s target. The player had also 50 ball touches per match, an incredible statistical factor for a goalkeeper.

January: Barcelona – 2010/2011
MVP: Lionel Messi
The phenomenal player from Argentina scored 6 goals, provided 3 assists and had a dribble accuracy of 91%.

February: Real Madrid – 2011/2012
MVP: Xabi Alonso
The Spanish midfielder had a 91% passing accuracy, 84% interception accuracy, 0 bookings and 2 assists.

March: Borussia Dortmund – 2011/2012
MVP: Mario Götze
The German Central Attacking Midfielder claimed 2 out 3 Man of the Match Awards by scoring 3 goals and providing 2 assists before his injury some days later.

April: Inter – 2009/2010
MVP: Maicon
The Brazilian Right Wing Back had 84% cross accuracy, while he scored 1 goal and provided 3 assists.

May: Manchester City – 2011/2012
MVP: David Silva
The Spanish Central Attacking Midfielder provided 6 assists, scored 1 goal while having 87% through ball accuracy. His contribution provided the team the highest score in the whole sample.

The most critical field position for a high score is the Central Attacking Midfielder (graph above). The player’s abilities to create clear chances and stop early attacks seem to be the most critical in order to present more attacking football and less defending.

Results and Discussion

The research above presented the best and the worst ranked champions using the introduced calculated ratio. The best ranked team in the Premier League the last 8 seasons was Manchester City, while for the Serie A this team was Juventus, for Primera Division it was Real Madrid and for Bundesliga it was Bayern Munich. Regarding the leagues, Bundesliga due to high attacking productivity and less competition ranks first followed by Primera Division. In the complete club comparison including all the champions in the sample, Manchester City had the highest score and Bayern Munich the highest consistency. Finally, the most critical field position was by far the Central Attacking Midfielder (CAM) with 4 MVP awards in the 9 including months.

Conclusion

This weighted calculated ratio compared the difference between attacking and defending efficiency of the league champions the previous 8 seasons in the 4 most distinguished European soccer leagues. Of course, there are plenty other ratios that can identify the form of a team in the season (Strength of Science, 2017). Yet, using that simple descriptive formula, the researcher succeeded to extract plenty valuable insights regarding the previous seasons in the top-level European championships.

You may find all the technical mechanics of the research here: https://goo.gl/6rbxHn


Bibliography

Statistics: How to, 2015. Inferential Statistics. [Online]
Available at: https://goo.gl/veKRHv

Strength of Science, 2017. Practical Application of the Acute: Chronic Ratio in Football. [Online]
Available at: https://goo.gl/U8gjgJ

Who Scored, 2017. MVP per Month. [Online]
Available at: https://goo.gl/Go7tnH


You may explore the raw data just below:

Energy and Persistence conquer all things according to Ben. Franklin. Statathlon for me combines my two lifelong passions: sports and science. Statathlon will be successful because all of us in the team are determined to seize every moment and every opportunity that either we see or create from scratch.

Site Footer