These days, sport is increasingly embracing predictive analytics. Broadcasters and sports data companies offer up a vast array of statistics that enhance understanding of an individual match and/or entire competition. This data can help managers, coaches, or even fans, analyse the behaviour of individual players, or the performance of an entire team, allowing them to adapt their overall coaching strategy or playbook according to the insights gathered.
Another example, the 'Hawk Eye' goal-line technology — which was at first controversial — is now broadly used in stadiums around the world and has already proved useful in some tricky goal-line situations. This discussion was somehow surprising, since technology is already ubiquitous in the modern sports and there has always been a strong connection between professional sports and the idea of data analytics. For example, the film 'Moneyball', based on a book published in 2003, describes how a baseball coach with a limited budget optimises his players’ potential on the basis of an in-depth analysis of their match statistics. Realising the value of the data he has accumulated, he sets up a team by recruiting players whose financial value has been underestimated and takes it to the play-offs two years’ running.
Around the world sports fans, athletes and spectators have been enjoying the Summer Olympic Games. And so, questions relating to how big data and statistical approaches to sports analysis are top of mind for many. When Rio 2016 is put into historical context it might be seen as the ‘Data Analysis Games’ with athletes and coaches increasingly working with big data to obtain never before seen insight. For example, in a Forbes article, contributor Bernard Marr details how the British rowing team, 'has increasingly ramped up their data-driven analytics with one primary aim – to make their boats go faster'. In the velodrome, Laura Trott thanked ‘the team at home’ including the team nutritionist and the data analysts after winning gold.
But is it actually possible to manipulate data in such a way that you would know the winner of a match, race or game before it even begins — or halfway through? If yes, would that destroy the thrill of watching Jason Kenny blow competitors away in the velodrome or take the edge off watching the Premier League’s new boys, Hull City topple Leicester on the opening day?
Using real-time analysis to predict the future
While the system that 'Moneyball' utilised has generally proved its value for post-event analysis, there’s a missing piece in terms of actions that need to be analysed in real-time vs. waiting until the end of the match, game, etc. For instance, betting websites need to update the data continually, in order to provide current prices for players or on the results of a game. With new analytic technologies such as Hadoop and Spark, associated with machine learning, it is now possible to predict very likely outcomes in real-time.
By combining historical data with real-time analysis, not only does the model make predictions based on new data arriving in real-time, but this new data also helps to enhance the underlying historical model, forming a continuous cycle of evaluation. Throughout the lifespan of the system, the relevance of the model’s predictions is incessantly improved.
Analytics applications for all areas of activity
Although it seems rather a complicated process to understand this kind of architecture and algorithm, the latest solutions for data integration are becoming more simple to use and designed with the layman’s user in mind. They also help easily upscale models created by data scientists (especially with tools such as the Spark Machine Learning Library), using the example of traditional scientific research practices: from the lab to the production line, the sample piece to the industrial process.
So many other sectors can draw inspiration from the way sports is embracing big data to address best practices that ultimately impact outcomes. For example, airlines which need their aircraft to be completely reliable, can predict future maintenance cycles by analysing vast amounts of data from sensors placed throughout the aeroplane, and transmitted in flight. In the field of distribution, the power of Amazon’s recommendations algorithms is well known. Finally, other sectors (insurance, banking, health, etc.) can also benefit from the contribution of predictive analysis, for issues associated with customer relations.
These days it is vital for all sectors, just as in the sporting world, to remain competitive and to move towards predictive analysis. The way to do it is to examine the existing situation in such a way that it delivers highly valuable information, backing up real-time strategic, decision-making.
So becoming data-driven is becoming more of an imperative for all businesses—whether in sports, banking, manufacturing, retail or healthcare. But in order for companies to truly make big data work for them most effectively, transparency and data governance are vital, especially when it comes to building consumer confidence and brand loyalty. Questions such as: on what data sources does the model rely? Where does the data come from? Is it accurate? Does it comply with current industry regulatory requirements?
These are key questions and will be pertinent to betting companies looking to set their odds on who will take home the gold in men’s basketball if the US men’s volleyball team will be able to snag the bronze, or which metrics an emerging small business should focus on in order to generate a profit. Whatever the context, its increasingly clear – big data is the name of the game!