Şirag Erkol
Arsène Wenger is the best manager in European soccer history. Red Auerbach is the best head coach in ABA/NBA history.
No need to argue. It’s science.
Well, we’re sure some of you are going to still argue the subject—if nothing else, it’s a lot of fun—but in a study published in the Journal of Complex Networks, researchers at the Luddy School of Informatics, Computing, and Engineering have analyzed two large datasets to build networks of pairwise interactions—one-on-one interactions—between head coaches in each sport that measure and rank their career performance. The result is CoachScore, which ranks the top 100 coaches among the five top-tier European soccer leagues—the English Premiere League, France’s Ligue 1, Germany’s Bundesliga, Italy’s Serie A, and Spain’s La Liga—and the American and National Basketball Associations.
Wenger, who spent 22 years at the helm of Arsenal in the EPL and 10 more in Ligue 1, took the top spot for European soccer coaches. He was followed by Sir Alex Ferguson (26 years with the EPL’s Manchester United), Jupp Heynckes (20-plus years in La Liga and Bundesliga), Carlo Ancelotti (20-plus years with stops in all five top-tier leagues), and Pep Guardiola (13-year career in La Liga, Bundesliga, and the EPL).
For ABA/NBA coaches, Auerbach’s legendary career, mostly with the Boston Celtics, garnered top honors, followed by longtime San Antonio Spurs head coach Gregg Popovich and Chicago Bulls-Los Angeles Lakers head coach Phil Jackson. Don Nelson, who coached five NBA teams and won a record 1,335 games, took the fourth spot, while Lenny Wilkens, who coached six NBA teams and the NBA championship with the Seattle SuperSonics in 1979, rounded out the top five.
Şirag Erkol, a Ph.D. candidate in complex systems and networks at the Luddy School, came up with the idea to use network science—the study of network representations of physical, biological, and social phenomena leading to predictive models of these phenomena—to rank coaches while watching the 2019 UEFA Champions League Final between EPL teams Liverpool and Tottenham Hotspur. Mauricio Pochettino (ranked 28) managed Tottenham, while Jurgen Klopp, who came in at ninth in the rankings, managed Liverpool.
“These are great coaches,” Erkol said. “You can compare them using trophies, but other than that, there isn’t much that you could really do. I realized you can use networks to do the analysis because network analysis creates an environment where you can easily do it. Data is data, and network analysis is built to draw out answers.”
Working with Associate Professor Filippo Radicchi, Erkol based his metrics on PageRank, an algorithm used by Google Search to rank web pages. Erkol took into account not only the number of wins or points a team earned in a season, but also who those wins/points came from.
“For example, if a Phil Jackson team beats a team coached by Gregg Popovich, it’s more important to the metric than Jackson beating a team coached by someone lower in our rankings,” Erkol said.
European soccer coaches were measured on total games, wins, losses, and ties, while ABA/NBA coaches were measured on games, wins, and losses. Other variables, such as talent on the teams, injuries, level of competition in each league from year-to-year, etc., were not taken into account because there is no way to quantify such variables in any consistent way. Career length of a coach, however, does impact the ranking for a simple reason.
“If you have a coach who has a long career, it’s an indication of how good the coach is,” Radicchi said. “Otherwise, he wouldn’t be allowed to coach for so long. But as long as you try to quantify performance using a single number as we did, you, by necessity, have to neglect some of the possible dimensions of performance at play.”
For instance, championships are only one facet of the equation. Jackson won 11 NBA titles, Auerbach won nine, and Popovich currently owns five championship rings, but Nelson never won a title, and Wilkens won just one. Auerbach also ranks 12th on the NBA’s all-time wins list and 13th in terms of winning percentage, and Nelson, despite being the winningest head coach in pro basketball history, is just 54th in winning percentage.
“Counting wins/trophies is not a good measure because it is hard to compare the level of leagues (in soccer) and the level of competition between eras,” Erkol said. “We propose a measure here that does not suffer from these limitations. We also propose a dynamic version of our measure, which can potentially be used to analyze the impact of coaches on teams under various conditions (e.g. handing a team in decline to a rising/declining coach), which we have not pursued to answer in this study and left as a future research question.”
Still, each coach’s work over time and vs. each other helped elevate Nelson to the top five despite the fact he never ranked as the top coach in any single year. Wilkens achieved that lofty height just once in his career per the metric.
“One concept that we try to stress in the paper is that whenever you’re ranking something, as long as you aren’t focusing on a single season or a championship where you can quantify performance in some objective way, such as counting points, there is not a single recipe that is going to work,” Radicchi said. “There are very successful coaches who have had very short careers. Or, they have been less successful in terms of trophies but enjoyed longer careers. This is a multi-dimensional subject, and it’s a matter of picking the dimension you want to use to measure it. It really is the result of your recipe, and it’s based on some opinion.”
The current rankings, especially for basketball, could soon change.
“If Popovich coaches for two or three more years, he could replace Red Auerbach on top of our rankings,” Erkol said.
The researchers are considering expanding their study to the NFL, especially thanks to the wealth of solid data about the league that already exists, and it allows the researchers to practice their theories and craft. Measuring, predicting, and understanding performance are scientific challenges that regard the study of any social system, such as social media, financial markets, professional sports, scientific innovation where there is competition for prestige, recognition, awards, social status, popularity, leadership, wealth, fame, etc.
“It’s just data,” Erkol said. “If the data for the NFL is anything like the NBA, it should be pretty easy to do because the data is so clean.”