Moneyball: Art vs Science

There is a constant debate between purists in sports and data professionals about whether the game is purely an art or can be measured through (data) science. Sports analytics is my passion, and while watching different sports, I am thinking about how data can be applied to this game.

A few years back, I had the privilege of hearing Mike Forde , the former director of operations at Chelsea Football Club , share his insights. Sports, as an industry, is unique; approximately 80% of total revenue is allocated to employee salaries, primarily players. Consequently, player recruitment stands as a pivotal aspect of any sport. To put the magnitude of this challenge into perspective, consider that the total value of footballer transfers in Europe amounts to a staggering 4 billion euros annually.

"We spend 80 percent of our revenue on three percent of our workforce - that's like having 30 executives on staff who have been with the company for an average of 1.9 years." - Mike Forde

Hence, the significance of ensuring the right players are enlisted cannot be overstated. In his speech, Forde delved into the ongoing debate regarding the relevance of Moneyball: is player selection an art, or can it be approached scientifically through data analysis? His response was a nuanced one, advocating for a synthesis of both methodologies, emphasizing the importance of posing the right questions to the data.

"99 percent of player acquisitions are about who you don't hire, and here, it's not enough to just have a gut feeling anymore." - Mike Forde

With over 2500 professional football games taking place in Europe annually and a multitude of players as potential choices, relying solely on scouts to evaluate talent becomes impractical. Herein lies the value of data; it enables clubs to assess individual players objectively.

In recent years, clubs have gained access to extensive datasets, but not always straightforward to interpret. While it may seem intuitive to base decisions on metrics like goal scoring or assists, such outcomes fail to encapsulate the essence of a player's contribution.

Chelsea's innovative approach involved employing data scientists with minimal football knowledge, tasking them with analyzing data from a fresh perspective. This endeavor led to the identification of novel Key Performance Indicators (KPIs) uncovering patterns previously overlooked. I always believe crowd-sourcing insights from people with diverse backgrounds will bring lateral thinking into your data-driven processes.

One such revelation pertained to Claude Makélélé, a French midfielder renowned for his tenure at Chelsea. Despite being the first choice of his coach in every lineup, conventional KPIs failed to highlight his significance. His name became synonymous with his position, i.e., Makélélé's role. His contributions transcended statistical measures; his defensive prowess and tireless work rate were critical to the team's success.

Introducing a new KPI, "Glory Hunter Yards vs. Hard Yards Runner," shed light on Makélélé's impact. What are these two terms? A glory hunter only runs when their team is in possession, and the "hard yards runner" is the one who runs when the opposition is in possession, e.g., during a counterattack and falling back to save the goal. This new KPI revealed that Makelele had the highest number of "hard yards" in the entire league. He was making the grinding fallback every time there was a counterattack. He was the workhorse of the team. He wasn't scoring goals or making crucial passes, but the people knew he was important. People thought this was the "art" of coaching because this guy's skills weren't visible in the data. The problem was that the team wasn't using the right KPIs.

So, the next time you think about whether you need an artist or a scientist to make the decision; consider whether you've employed lateral thinking to craft the appropriate KPIs. After all, true understanding often lies beyond conventional metrics, waiting to be revealed through innovative analysis.

If you like, please subscribe to the FAQ on Data newsletter and/or follow Fawad Qureshi on LinkedIn.