The first round is over, we have now seen each nation’s first foray into the 2018 FIFA World Cup. Far from giving us clarity, the intriguing the question of ‘who will win the competition?’ is perhaps even harder to predict than it was before the tournament started. The favourite nations have failed to convince, with only Ronaldo of the pre-tournament superstars has performed to a level worthy of the title.
Before everyone abandons their predictions, we take a look at three possible predictive models that might help you stick or twist.
Candidate 1: FIFA world ranking
The first – and most obvious is the FIFA world rankings. One might reasonably assume number one in the world equates to the greatest likelihood of success.
On these grounds, the most obvious winner is Germany with Brazil as runners up. This might seem like a relatively solid bet on the surface; Germany after all, are defending champions and, despite an opening fixture defeat against a good Mexican team, have been in fine form in the intervening years. Only beaten by France in the Euro 2016 semi-finals and storming their World Cup qualification group with 10 wins from 10 games.
Brazil, meanwhile, has been undergoing a revamp since their 7-1 humiliation by the Germans on home soil in the 2014 World Cup. Subsequently, they have changed manager and won the notoriously difficult South Americas qualification group by 10 points. Despite only a point vs the now infamous giant killers Iceland, they have the attacking strength to test any team in the world.
Over the past 4 tournaments of the teams that have entered the World Cup carrying the number one ranking, no team has gone on to win. Indeed, the team ranked second has won in three out of the four tournaments (2002, 2010 & 2014), whilst Italy was not even within the top 10 of the world rankings (yet received a top-eight seeding for the draw) when they triumphed in 2006. Perhaps top-spot comes hand-in-hand with the weight of expectation?
Candidate 2: A more thought-out ranking?
Opta proposes an alternative prediction model. A team’s attacking and defensive strengths are factored from previous results, in conjunction with ‘World Cup specific variables’ to predict the outcome of each fixture. These include elements such as prospective fixture (and path) difficulty, home-field advantage for the tournament and previous world cup success. In their model, Brazil is predicted as the most likely winners, albeit with a mere 14.2 % chance of actually lifting the trophy. Germany, Argentina, Spain and France then jostle at around a 10% chance of glory.
However, once again the model is not able to factor how changes in personnel might affect a team. Opta themselves state, for example, that if Messi were to be injured Argentina’s chances of success would plummet.
Candidate 3: Competitiveness of Domestic Top-Tier League:
Finally, a paper in 2005 from Halicioglu used a statistical model to argue that the success of a nation’s team is related to the degree of competition in their top-tier domestic league. Arguing that a nation’s domestic league is the primary breeding ground for new talent (even if this talent subsequently moves abroad), the author posits that nations whose players have been exposed to more competitive domestic football will inherently attain a national squad with greater talent than one whose league is relatively less competitive. Essentially, the more unpredictable a nation’s domestic league champion, the more competitive that nation’s international team is likely to be. The author validated the model against the Euro 2000 tournament outcome, finding that France (the eventual winners) also had the least predictable domestic league at the time.
For those backing England, this model is the one to use to justify your choice. In the last 5 years, the English Premier League has seen 3 different winners and 5 different runners up. Among the tournament favourites, only the Brazilian league can match that competitiveness over the same time period.
Whilst the model appears to have good general predictive strength, it does not factor for the number of ‘home-nation’ eligible players within a domestic league. This may be an interesting moderator to add to the predictive model.
Who will win the World Cup?
The various factors contributing to team-sports make it near impossible to predict with any degree of certainty how a team might progress. Indeed, Halicioglu (2005) concedes that football is a highly volatile and unpredictable sporting environment. Measurable factors can be used to compute likelihood, whilst non-measurable factors then disrupt the prediction.
Looking at these three predictive models and the stats that surround them, come the 15th of July, a hard won point against Switzerland will be a distant memory as Brazil lift the World Cup…
Sport Science Agency works with brands, broadcasters, rights holders and agencies to create insight, experiences and content from the latest sport science research. If you want to know more just drop us a note via firstname.lastname@example.org and we can arrange to go for a healthy performance boosting drink.