Decoding the football world cup through data analytics
The maximum degree of Magnus force possible for different players against a free-kick wall can help improve the odds of the right prediction and may indeed impact team composition and style against specific opponents
The world has been in the midst of World Cup football fever for the last few weeks. In India, late-night games with penalty shootouts have led to many groggy workers in offices. Some big-name countries like Italy did not even make it to this World Cup (widely blamed on coach Giampiero Ventura) while others, like Germany (which lost to South Korea and Argentina), have already been knocked out.
The low frequency of goals and history of “upsets” has led to the thinking that football, unlike many other sports, cannot really be deconstructed analytically with the help of data. Sports such as baseball, American football, basketball and cricket are now being studied analytically and deeply for teams to gain a competitive advantage.
The 2011 movie Moneyball was based on a book that discussed an analytical system used in baseball by the Oakland Athletics to assemble a top-notch team on a limited budget. Firms like Cricket-21 provide detailed match analysis, including data, graphics, real-time video clippings, and analysis of the opposition. It is because of techniques like this that India’s Suresh Raina gets bounced out on his ribs and the opposition bowls full and straight (an in-swinging delivery if possible) to Australian Shane Watson. Until recently, football had escaped this data-oriented approach.
First, some history for context. The inaugural football World Cup was held in 1930, and won by hosts Uruguay, who beat their neighbours, Argentina, 4-2. Brazil have won the most number of World Cups (5) followed by Germany and Italy (4).
In recent times, the average number of goals per World Cup match has been in the range of 2.2-3.0. It used to be between 3-4.5 in the 1950s.
As defence and passing have become more systematic, and fitness a more general requirement, the number of goals has gone down. Miroslav Klose of Germany, with 16 goals, and Ronaldo of Brazil with 15, lead the individual tally for goals in World Cups. The legendary Zinedine Zidane of France leads with the most cards. These aggregate statistics, though, are of little use in predicting the outcome of a specific game.
Unlike club football, World Cup football analytics is further complicated by “a small data set” problem since it is played once in four years.
Individual players can represent their countries in multiple World Cups—as recent examples, Thierry Henry of France and Xavi of Spain have each represented their countries four times. Despite this, country teams are put together mere weeks before a World Cup and disbanded soon after because players have strict club commitments.
When winning or losing is guided by a low-frequency event, statistically speaking, it is difficult to distinguish prediction from luck. However, high-frequency statistics like the number and length of passes, the zone occupation on the field relative to the opposition’s ability to defend the zone, the maximum degree of bend possible for different players against a free-kick “wall” and other such factors can help improve the odds of the right prediction and may, indeed, impact team composition and style against specific opponents.
When a football is kicked in the air with spin, it is subjected to the “Magnus force”, otherwise called the bend, that is proportional to the ball’s angular velocity and the distance of the kick and inversely proportional to its mass. Even the origin location of the spin on the ball affects its trajectory—Philippe Coutinho of Brazil provides it from the instep and Kevin De Bruyne of Belgium glances it from the outstep. Cristiano Ronaldo kicks the ball with the perfect “drag transition”—a technique where the ball rises over the wall but dips into the net. If you want to hit a straight shot at goal without spin, you have to make contact with the ball for the least amount of time and the best way to do that is by hitting its valve. Of course, players do this from practice and muscle memory, but, nowadays, coaches and handlers are using analytics and physics to help them along.
Beyond the use of data for team management and competitive advantage, analytics is also being used for fan engagement. This is a trickier proposition as data and inference is being delivered with the primary objective of more sustainable emotional interaction. Yet it is becoming necessary to use these techniques to create online communities, increase interactive engagement and build digital trust.
It is often said that football is a metaphor for life. Traditionally, it has taught us the importance of a team, the value of picking oneself up after a fall, and of transferring the ball to the best positioned member of the team (paradoxically often the result of a cross). Data analytics in football additionally teaches us that life, though complex, can be broken down into more predictable and repeatable chunks and that practice will allow for “bending” otherwise ordinary expectations. Methodicalness and discipline aid one’s ability to go through life and overcome opposition. And last but not least, given the statistical chance element, a miracle is always possible.
P.S.: “Success is no accident. It is hard work, perseverance, learning, studying, sacrifice and, most of all, love of what you are doing or learning to do,” said the legendary Pelé, regarded by many as the greatest footballer of all time.
Narayan Ramachandran is chairman, InKlude Labs. Read his earlier columns at www.livemint.com/avisiblehand
Comments are welcome at firstname.lastname@example.org
Editor's Picks »
- 5 issues that’ll dominate RBI board meeting tomorrow
- Future Retail’s Q2 result shows improvement in same-store sales
- Private insurance firms grow at the expense of LIC stuck with a sick bank
- Page Industries’s lofty valuations get a reality check in Q2
- Q2 results: Grasim’s Vodafone Idea stake is proving costly