The finale of 2019 season is upon us and that means we are one year removed from another decade into the 21st century. Everything on earth is in a downfall except the quality of football and brunch places. This is the first edition of ranking the best at every position in the English Premier League (EPL).
Premier league started employing dedicated teams to chart advanced data from each game and every player involved starting 2007-08. The availability of advanced statistics have led to a system which can help us predict the future and appreciate the past, efficiently.
Why is ranking the best talents from the past based on statistics and numbers important? The answer to the question starts with another question, Is the current state of data analytics in sports good/ trustworthy? And the answer is no. The value of sabermetric in football lies in improving the scouting of future generation by separating relevant data from irrelevant ones. The concept is deep seeded in the data collected which can be verified by observations and past experiences. One way to determine the relevancy of a statistical data is by studying their impact on past footballers, then comparing the analytics based grading with the conventional wisdom of the pundits based on film study and eye test.
Last year, I worked on developing a system based on the player earnings to predict how much money a player is worth relative to their production. I called it the Soccer Investment Grading. The tool can be useful to evaluate players overlapping the same timeline and negotiate player contracts based on player production and their respective market value (It does require a lot of improvement). The new formula disregards the complication of money and attempts to grade the stars of the past based on comparative skills across all eras.
During the grading of best central midfielders in English Premier League (2010-2019), we do not consider players who spent most of their career as defensive midfielder. I shortlisted a list of 9 deserving midfielders based on public perception from the present decade for the study. Three players from Chelsea made it in the top 9 - Eden Hazard, Juan Mata (Chelsea & Manchester United) and Cesc Fabregas (Chelsea & Arsenal). Cesc Fabregas had an underwhelming finish to his career in England but had a very dominant presence in world football during his first few seasons with Chelsea and Arsenal in 2010-11. Mesut Ozil and Steven Gerrard made the list from Arsenal and Liverpool respectively. Christian Eriksen was the lone representation from Tottenham alongside Gylfi Sigurdsson (Tottenham, Everton and Swansea). The list was rounded by representation from Manchester City’s David Silva and Yaya Toure.
The framework is based on 13 independent variables related to midfielder production. Each variable (such as assist per game) is given a multiplicative factor in the formula based on their influence towards outcome of the game. The positive variables are added (such as successful dribbles per game) and negative variables get subtracted (like bad control per game) to assign the final grade.
The number of assists were given the most value followed by goals per game and key passes per game respectively. Through ball, accurate long balls and successful dribbles were also considered as important traits in grading the players. Defensive influence of the midfielders were evaluated based on successful tackles, interceptions and bad control. Big chances created per game was also a notable variable in rating the players. ‘Big chance created’ charted the play where the midfielders action lead to one on one situations or created scoring chances inside the penalty box. All the raw data were collected from offical website of the premier league and whoscored.com.
Steven Gerrard and Cesc Fabregas had the highest accuracy in long balls per game at 6.62 and 4.15 respectively. Their long ball production could be correlated to the deeper positioning in the midfield as a conventional no. 8 instead of no. 10. The successful dribbles per game could be directly equated with bad control and loss of possession per game. Hazard had the most successful dribble of 3.69 per game and was dispossessed 2.12 times every game too (highest among test players). Steven Gerrard had the lowest successful dribbles per game as well as the lowest rate of bad control (except when he slipped) and dispossession, since 2010.
The final grading awarded Gylfi Sigurdsson with the lowest grade of 12 whereas Cesc Fabregas was the unexpected leader with a grade of 25. Fabregas had the highest through ball (1.36 per game), accurate long balls ( 4.15 per game) and assists (0.32 per game) among the 9 players over the last 10 years. The midfield data evaluation revealed that Eden Hazard did not grade positively as a central/ attacking midfielder and his value lies in a more advanced role, as a winger. I expected Hazard to grade the highest but he got a below average 20 with very high stat lines in goals per game and successful dribbles (statistic more synonymous with forwards). Appearances were given a number as well with the highest worth the most at 1 and then decreasing by a rate of 0.05 to reward their longevity. David Silva had the most appearances in the past 9 years with above average stat line in every relevant midfield category.
Christian Eriksen had the 3rd lowest score of 21 ahead of Juan Mata (2nd lowest) with commendable assist per game of 0.29. Eriksen was tied at the top with 2.67 key passes per game alongside David Silva and Mesut Ozil. Yaya Toure and David Silva had top tier grading in every criteria, making them the most complete midfielders in the group since 2010. The study gave a mean grade of 22 and a mode of 23 (Silva, Gerrard, Yaya Toure). Even though David Silva and Steven Gerrard got the same grading, Silva would be ranked ahead of Gerrard because of his superior statistics in more relevant variables (for instance, assist per game is more important than accurate long balls per game).
Instead of ranking the players, the appropriate methodology would be to sort the players into tiers. Eden Hazard would be in Tier - ‘Not a midfielder but a forward’. David Silva, Yaya Toure & Cesc Fabregas would be in Tier I. Gerrard & Ozil would constitute Tier II and Gylfi Sigurdsson, Juan Mata & Christian Eriksen would round up Tier III.