The first ball of FIFA World Cup 2014 was kicked in the Arena de Sao Paolo, Brazil on June 12 as Brazil went head to head with Croatia in the competition’s opening game.
The qualification process, involving 239 nations, began as far back as 2011, and now just 32 teams remain to battle it out in the finals. World Cup hosts Brazil are hotly tipped to lift the trophy, but when the pressure is on, anything can happen in tournament football.
All 32 teams have spent the past four years building up to this moment – spending hours preparing both physically and tactically, as any extra competitive edge could make all the difference. It’s little wonder then that so many football teams are turning to big data analytics to find just that.
As the old saying goes, ‘hard work beats talent, but only when talent doesn’t work hard’. It used to be the case that you could win by either being naturally gifted, or by working harder than the rest, recalls Alex Philips, senior managing consultant, Business Analytics & Optimisation, IBM.
"Now, talent identification programmes are so strong, and training regimes so finely tuned, that, at the very top level, everyone is talented and everyone works hard," he says. "So to find that extra edge in performance you need to look elsewhere – data analytics can tell you where that extra edge might be, and help you grab it before your opponent."
In situations like the World Cup, ‘big data’ is so much more than mere business jargon. Here, it could be the difference between winning and losing. And although big data is a term that has been heavily marketed in recent years, it’s certainly not new to the world of football, as Chris Anderson, author of The Numbers Game, explained at this year’s MIT Sloan Sports Analytics conference.
Content from our partners
"Some people have described what’s going on with soccer analytics as a revolution but I think it’s a bit of a misnomer," the quantitative football analytics pioneer and former semi-pro footballer commented.
"I think it’s been much more of an evolution. Very few people know that soccer is really the original hotbed of analytics."
Anderson was referring to the work of Charles Reep, an accountant and RAF wing commander, who back in 1950 began attending English football matches and making notes about the action that unfolded on the pitch.
"He developed a coding system," Anderson explained. "He sat down at matches for 60 years and noted with his own system all the events on the pitch and collected reams of data that he then put into practice, working with clubs through the 1960s to 1980s in England."
He was the original soccer analyst – an analytics driving force. "That history and Reeves’ successes and failures with teams has influenced how all this has evolved and reverberated through clubs over the years," said Anderson.
The history of analytics, particularly in football, has been stop/start over the decades. Right now, though, we’re again in a period where analytics is expanding and evolving and, now, two major sets of big data analytics solutions exist today in football, says Duncan Ross, director of data science, Teradata UK: Those about individual athletes and those about teams.
"There is more information than ever available on the individual, even for a club runner," he notes. "But professional athletes have much more sophisticated data including blood oxygen levels, respiration and many other metrics. A vast amount of individual information is building up into a knowledge base of how the body works."
Analysis of teams is already widely in use and works very well in sports like baseball where there is a lot of data regarding individual performances, but it is more challenging in a team environment such as football.
"Using data captured during a match is difficult and challenging from a mathematical perspective because there are so many different elements," says Ross.
Delving further into the analytics solutions currently available, there are a variety that encompass injury prediction and management, team selection, training regimes, and scouting for new players, Philips says. Such solutions make use of many types of data including heart rate, player position, speed and acceleration, and even the forces received by the body during tackles.
Some teams are analysing this data in near real-time during training, as well as looking at post-match and historical data to further improve training and match tactics, Philips says, although the analytics behind these solutions can often be very complex and difficult to get right, especially for things like injury prediction.
"Cycling pioneered much of the analysis we see now, but other sports have followed in its wake," he adds. "Dave Brailsford, former performance director of British Cycling, said that it was important to understand the ‘aggregation of marginal gains’ – in other words, making small improvements to many aspects of what the individuals and team do adds up to make a significant difference. Analytics is key to enabling this approach."
There are also a lot of exciting analytics developments taking place, with the cost of data decreasing and, perhaps most crucially, data becoming more readily available.
Laurie Miles, head of analytics, SAS UK & Ireland, comments: "What’s really improved is the accessibility that sports teams now have to advanced analytics, and how answers to very specific questions can be delivered in minutes or seconds and in a format that’s easy to understand.
"The results can now be easily shared over multiple mobile devices so that even coaches out on the field can get immediate access to the results they’re after."
We are also seeing a move away from just pre- and post-match analysis towards an in-match analytics focus.
Steve Houston, an expert technical scout previously with Chelsea FC and Hamburger SV, points to Formula 1 as a prime example.
"F1 has the best data modelling solutions from across all sport," Houston says. "In 2008 when Lewis Hamilton won the championship he went into the last race needing to finish fifth. On the final lap he was in fifth place but he had Vettel behind him competing for that place.
"The McLaren guys were bringing in data from their own cars, competitors’ cars, the weather, historical data, and it was all being fed into a model. And that model with one lap to go told Hamilton to let Vettel through. The analytics predicted that they would both catch the guy in front within two seconds as he was using the wrong tyres. The principal of the team trusted the data and made that call. Hamilton won the championship."
It’s unlikely that all of the World Cup competitors will be making use of big data analytics but plenty will be. However, exactly what kind of impact will this have on the tournament’s outcome remains to be seen.
Philips says: "International football is a complicated beast because, for the players at least, it’s something of a part time job. This is challenging for the international coaching teams because they don’t have direct influence over a player’s day to day training regime and habits."
Football as a whole still has a way to progress though, thinks Skytree founder and CEO (and ex-Bayern Munich player) Martin Hack, part of which is a reluctance to introduce technology due to a deep respect for tradition.
"Some people are reluctant to apply new technologies unless you actually prove to them the value," he says, prediciting that we will see an underdog triumph similar to ‘Moneyball’ heroes the Oakland A’s ‘within the next few years’
"The value of data analytics has already been proven, its not like black magic," he believes.
However, there are indications that analytics is playing an increasingly important role in team and squad selection, and not just with respect to pure performance or ability. The France coach openly admitted that Samir Nasri is one of the best 23 players in France, but chose to leave him out of the squad because of the effect he has on the overall team dynamic if he’s not in the starting XI.
So while the World Cup rages this summer, what, if anything can businesses learn from how the world’s elite footballing nations utilise big data analytics?
There are some low-hanging fruits for business in transferring what has been learned in the sporting world, according to Ross. "One area is that of handling employees and their concerns about the use of measurements and metrics. If handled badly it can cause as much damage as good.
"Think about how, in many organisations, the people front-of-house, who have the most contact with customers, have the least training."
But companies also need to learn how to analyse and react quickly to data, as football coaches might during games, Philips adds.
"Many businesses struggle with information overload, to the extent that they can’t separate the signal from the noise quickly enough," he says. "What they need to do is understand the crucial pieces of information that they need to make day to day decisions and interventions and focus on making them available to decision makers as and when they need them – not when the next reporting cycle comes around."
The big data market in businesses is actually very mature – there are many technologies available that have the capability to handle the volume, variety and velocity of incoming data. This causes a certain amount of paralysis for anyone starting on a big data journey, though.
Big data projects often come with big commitments and the amount of choice available means that people are sometimes so nervous about choosing the wrong solution that they don’t choose at all.
Philips says: "The challenge that a sports organisation, or indeed any other organisation, faces is selecting the right technology for their particular problem. To that extent they need to consider what the desired outcome of the big data analytics is and what they want to be able to do differently once that outcome is delivered."
Once they have done that they will be able to decide whether they need an option that will enable them to analyse vast quantities of data in real time, such as Infosphere Streams, or the capability to handle structured, semi-structured or unstructured data, such as a Hadoop based system.
The true value of analytics in sports is yet to be seen, thinks Hack, but this will soon change. "Five to ten years ago nobody really cared about analytics, he says, "but now they can give people insights and capabilities that just weren’t there before.
"If you are an industry leader, you’ll do anything to keep that edge – if that means you analyse more data in a smarter way, then that’s a good thing to do, as it gives you better insights."
So when the Jules Rimet trophy is awarded to the winning nation on July 13, the victory may owe more than a little to the world of data analytics. And with the market set to grow exponentially, it may be that we look back in future years to Brazil 2014 as a landmark moment not just for the technology industry, but for sport as a whole.
To find out more, check out the latest edition of the CBR app – available for all iOS devices from the link above.
