The FIFA World Cup Russia 2018 is being hailed as one of the most exciting yet thanks to a catalogue of shock exits and unpredictable results.

With goals and talking points galore, Germany exiting the group stage after surprise losses to Mexico and South Korea and much-fancied Spain tasting last 16 defeat on penalties to rank outsiders Russia, it’s no wonder #insane has been trending on Twitter.


But what if this mother of all footballing spectacles is not as unpredictable as first thought? What if Friday and Saturday’s quarter-final line-up was, in fact, quite predictable?


Kitman Labs, the award-winning sports science technology company, believes when it comes to predicting the future, there is no better indicator than past data. It is a tried and tested method that Kitman Labs, voted BT Sports Industry Sports Performance Technology of the Year 2018, harnesses in helping leading sports clubs around the globe better understand player data to drive performance and reduce the risk of injury.


And by drilling down into the data from the past five World Cups, Kitman Labs has unearthed a statistical model that accurately foretold three quarters of the nations who qualified for the last eight.



For a ‘Who will win the World Cup? By the numbers’ report, Kitman Labs studied results at France 1998, Japan/South Korea 2002, Germany 2006, South Africa 2010 and Brazil 2014. They discovered that each team’s goal difference during the group stage is an accurate yardstick for how far they will progress in the tournament.


On average, teams who reach the quarter-finals emerge from the group stage with a goal difference of four-five. Russia, Brazil (both plus four), England and Uruguay (both plus five) all finished within this bracket, while Croatia (plus six) and Belgium (plus seven) qualified even more impressively. So, in theory, those six teams’ passages should have come as no surprise. It just may not have felt that way when Russia, Croatia and England fans were watching their heroes negotiate nerve-wracking penalty shootouts! Sweden (plus three) and France (plus two) are the anomalies to this model.



Kitman Labs’ model predicts the teams who finished the group stage with the greater goal difference will make the last four. This suggests Uruguay will shock France and face Belgium, who are destined to stun favourites Brazil. In the other side of the draw, Croatia are tipped to end hosts Russia’s dreams and set up a semi-final date with England, who will overcome Sweden.



Using this group stage model, Kitman Labs predicts a third World Cup victory, and a first for almost 70 years, for Uruguay. Croatia, as first-time finalists, will have been vanquished. The 3rd/4th Place Play-Off will see Belgium overcome England for the second time this tournament.


These finishing positions are not just based on group stage goal difference, but also on goals conceded. All four teams finished with a group stage goal difference of at least plus five, which matches the average of the previous five World Cup winners, but it is teams who conceded fewer times that, statistically speaking, enjoy the most success – Uruguay (0), Croatia (1), Belgium (2) and England (3).


Yet in their report, Kitman Labs also highlight team performance in the semi-finals can create its own unique statistical model, with the potential again to change the outcome based on goal difference.


From this model, Kitman Labs believe the winner of the final on July 15 can be predicted based on semi-final results. Over the past five years, the World Champions have won their semi-final by an average of 3-0, with the runners-up averaging a goal difference of 1-1 – which possibly spells more penalties!


Either way, their predictions are certainly interesting…and by Saturday evening we’ll get to see how accurate Kitman Labs’ model continues to prove at this most unpredictable of tournaments!


Kitman Labs works with sports clubs across the globe to help them understand data and make better decisions to optimise the health and performance of their athletes.  

For further information, and to read the full report: