The real Euros battle lies with predictive analytics

Accuracy rankings as of June 25th 2021

There’s excitement at the DeeperThanBlue offices in Victoria Quays. With Euro 2020 now in full swing, Head of Analytics Simon Harrison is giving bookkeepers a run for their money in the form of predictive analytics and human guesswork.

Euro 2020 comprises of 51 matches between 24 countries, split into six groups. Before the start of the first round, we asked people to submit a full set of score predictions for each match in the group stages and collated them in our database, segmented by each person’s knowledge level.

We also took some of the world’s best algorithms and popular publicly shared predictions and benchmarked them against our human predictions to see which will come out on top – a hard fight to win when you’re up against the likes of William Hill, Ladbrokes and the FIFA World Rankings.

We also created our own prediction algorithm, which uses a wide range of sources, historic results, squad values and recent performance to predict the most likely score outcome.

All this ties together nicely into our own Power BI dashboard, which presents who will win based on the data we’ve collected – both human and AI – as well as highlighting the most popular scores.

Harrison said, “We love doing this for the major football tournaments it is always a lot of fun. It’s a good chance to put our driver analysis skills to the test in notoriously difficult to predict football matches as a change to the usual business-oriented applications.

“So far in the first round the human predictors have the edge, but we will see that as more games are played, the consistency of the algorithms will prevail.

“Applying this to exciting football tournaments is the only time that my 14-year-old son thinks that what we do is actually cool.”

• Share