About Petter & Nils

Petter Rudi Petter is responsible for match analysis for Molde Fotballklubb – where he analyses own team and upcoming opponents. Prior to this, he worked as scout and player developer in the same club. Between the age of 17 and 34, Petter played at the highest levels in Norway, Belgium, Italy, England and Austria, as well as for Norway (57 matches for age specific teams and 46 matches for the senior national team). He has UEFA A-License and is currently studying statistics to improve his analysis. Nils Rudi Nils is Associate Professor at INSEAD and has a PhD from Wharton School of Business. His research focuses on analytics in two disciplines: supply chain management and football. He is currently involved in two start-ups: Bravo Lucy (analytics for supply chain management and retail) and analy7ics (analytics for football).

ManU 66% – ManC 34%

Aside

The key football question of the day is: which team will win the Premier League? Some say Manchester United. Some say Manchester City. But this is not really the relevant question to ask. The right (although less tabloid) question is: What are the respective chances of the two teams?

We have spent significant effort in developing software that will calculate the probabilities of the various positions at the end of the season. Below we’ll give results for the probabilities of who will win the Premier League. But first, let’s look at how the software works.

The first step is to calculate the probabilities for the various outcomes. This goes beyond which team wins (or if there is a tie), as goal difference also plays a role in the ranking. These calculations are based on odds and statistical analysis of the differences between each team’s strengths (attack and defence) based on previous matches (where more recent ones count more than those some time ago). Based on these calculations, a computer will “play” the remaining matches which results in a “final” rank. This is repeated 10,000 times, and the probability for each team’s final rank is based on how frequently it occurs in the 10,000 rankings. For example if ManU ranks first 6,603 out of the 10,000 times, this is rounded to a probability of 66.0% that ManU will win the Premier League.

But when matches are in play, things get more complicated, as one needs to take into account remaining time in each match as well as current score. This is done by combining the method above with an advanced algorithm. (An algorithm is a series of mathematical operations which, together, calculate a result.)

So let’s look at some of the results from the calculations.

As of now (Monday afternoon), Manchester United has 66.0% probability of winning the Premier League, while Manchester City has the remaining 34.0%.

Now let’s then take the calculations one step further and consider how the situation will look contingent on various results in tonight’s match. If City wins 1-0, then the current situation will be reversed and City will have a 66.7% probability of winning the league. This is in part due to City’s superior goal difference outweighing the fact that they have a difficult match away against Newcastle. However, if the match ends in a draw, then ManU will have an 84.5% probability of winning the league. And the scenario of ManU winning 0-1 pretty much seals it, as ManU will then have a 99.2% probability of winning the Premier League.

It’s also interesting to consider the probabilities for who will win the league based on some possible half time results:

  • City leads with one goal at half time: 52.3% probability that City wins (47.7% ManU).
  • Tied at half time: 70.4% probability that ManU wins (29.6% City).
  • ManU leads with one goal at half time: 88.5% probability that ManU wins (11.5% City).

A few things are certain, however. A team from Manchester will win the Premier League. And tonight, much beer will be drunk across that divided city.

To end off, we include an example of percentage calculations from Manchester.

Petter & Nils