Although long-time fans of a certain generation might still feel that the Premier League is the new kid on the block, it’s been well over three decades since its inception, and in that time, a huge amount has changed, not least in terms of how football predictions are made. Early on, print media ruled the roost, both in terms of where data was collated and published, and how the conversation around the sport was consumed and controlled. Smash cut to the modern iteration of the game, and the digital takeover is total, demonstrating just how much the Premier League has influenced, and been influenced by, the wider market.
The Prestige of the Press
When sports pages were the main source of football stats and discussion, the majority of predictions provided were subjective, as individual pundits would cast their eye over recent scorelines and teamsheets to make assertions about how upcoming matches would play out. In amongst this, the usual mix of biases and a tendency to adhere more to media narratives than on-the-pitch realities meant that while the press coverage might have been entertaining to read, it was not particularly accurate in its prediction of Premier League games.
Part of the problem was a lack of depth in the data itself. League tables and top-level scoreline statistics were about as much as pundits and punters had to go on. Compared with what you’ll find on a modern EPL odds site, it was a comparative desert. But that would all change.
The Statistical Revolution
Starting in the 2010s, there was a real sea change in how football analytics was handled, and an entire analytics industry blossomed around the Premier League in particular, with clubs themselves employing companies to track a range of previously unrecorded data points, looking for contextual indicators that give a lot more insight into performance than scorelines alone.
This is when we got things like the emergence of expected goals, as a measure of the quality of goalscoring chances according to factors like shot location and assist type, along with positional and spatial data to tap into the intricacies of ball movement, rather than possession percentages being the main metric used.
The Adoption of Algorithms & Automation
With a deluge of data now collected across every single Premier League match, it was equally important for analysts to be able to process all this information and extrapolate meaningful conclusions, and that’s where algorithms and machine learning entered the equation, bringing us up to speed with where things stand today.
In addition to being able to interpret historical data more precisely as a means of predicting future outcomes, we’re now in an era where it’s possible to run full-scale simulations of games, a little like how meteorologists simulate the weather for long-term forecasting. And it was only because the Premier League made these advancements financially viable to pursue that they have even been developed, so it really is responsible for much of how football prediction technology has turned out in the 21st century.