The World Cup group stage: predictions through the betting markets

For the rest of us who are already converted, this is pure bliss. We get to watch national teams play football and discuss/predict/yell at/celebrate (choose any) games and tactics. The competition starts off with a group stage in which the qualified teams are subdivided into eight groups of four teams each.

Each team plays against every other team in the group and is awarded three points for a win and one point for a draw. These points are the key criteria used to rank teams such that the top two teams per group qualify for the second round. In this article, I try to make some group stage predictions based on the expected outcomes as implied by betting odds as of Monday June 2.

The betting market has not yet fully settled for all games and the most likely score could only be obtained for the first 17 games. For example betting on Brazil winning 2-0 in their opening match against Croatia would yield the lowest return from all scores, making it the most likely score. In this game, betting markets are implying that Brazil have a 75% chance to win the match while Croatia have a 7% chance of winning. The remaining 18% relate to the probability of the game ending in a draw.

My predictions for the remaining 31 matches are based on the following algorithm

• If the probability of a draw is more than 30%, then the game is assigned as a draw;
• Else if the favourite has a probability of more than 75% of winning, the team is predicted to win by a goal difference of 3;
• Else if the favourite has a probability of winning between 60% and 75%, the team is predicted win by a goal difference of 2;
• Else the favourite is said to win by a goal difference of 1.

As an example, consider that the match Japan vs Colombia has a 30.4% chance of a draw and was predicted to end 1-1. Furthermore, the score for Algeria vs Russia was predicted as 0-1 since there is less than 30% chance of a draw and Russia has a 53% chance of winning. This simplistic procedure loosely follows a paper I jointly published last year. However in this case, the number of goals scored is a subjective decision based on my own thoughts rather than any objective criteria.

I also made three exceptions to the above:

• Uruguay vs England: The probability of a draw is less than 30% but the teams are considered very even in betting markets. Therefore I opted for a draw in this match.
• Croatia vs Mexico:  Although betting markets indicate that Croatia is more likely to win than Mexico by a slight margin, I set Mexico as winners only because my seven year old nephew is, for some unknown reason, siding with them. He did advise me that they are a very good team.
• Honduras vs Switzerland: Using the decision rule above, I should have set Switzerland to win by one goal. However I think Honduras is more likely to lose their last game by a large margin than any other team in the competition.

These predictions form my submission for a free World Cup prediction competition in which participants attempt to guess the outcome of group stage matches. Participation is free but the top ranked individuals can win real prizes, including £80/€100 cash as first prize – it takes less than 10 minutes to participate! You can also follow updates on Facebook.

In a later article to be published on Thursday, I will discuss the betting market’s expectation of the second and final rounds as well as making a cheeky prediction. In the meantime, here is my predictions for the 48 group stage games.