World Cup prediction results (and some statistics)

Well, the World Cup has been and gone. Enthralling as always, this time the tournament proved to be more of a tactical battle. Although I felt Germany deserved to win, Spain did manage to stamp their authority and lose their ‘perennial underachiever’ status.

Let’s look back at my prior predictions to compare them to what really happened…



1. Team that completely meets expectations

Prediction: Democratic People’s Republic of Korea (North Korea)

No surprise that I was spot-on here. Although they showed some promise after losing very modestly to Brazil 2-1, they lost all 3 group games and managed to get drubbed 7-0 by Portugal. This took their overall goal difference to -11, cementing their place as the worst performing team in South Africa.

2. Average number of corner kicks per game

Prediction: 11

As I mentioned, the expected value of the number of corners in a game of football is around 11. The average number of corners per game in the World Cup turned out to be 9.875. Not far off. But is this statistically different from 11? I’ll do some hypothesis testing later on to find out…

3. Most boring match

Prediction: Slovenia vs Algeria

This one is subjective, but since I watched practically every match, I’d say I was right here. Slovenia vs Algeria was dismal. It finished 1-0, which doesn’t sound too bad, but believe me, nothing of note happened in the match whatsoever.

4. Manager who will make the worst decision(s)

Prediction: Diego Maradona (Argentina)

It could be argued that Argentina were undone by a smart German team partly because they stuck to their attacking, open nature, and this allowed them to be picked apart with a bit of planning. However, Maradona actually had a decent World Cup in that he didn’t make any shockingly bad decisions.

In fact, most managers did okay – although special mention must go to Raymond Domenech of France for helping his team along in its self-destruction.

5. Number of teams who will exit the tournament due to a poor refereeing decision

Prediction: At least 1

Whether a team exited the Cup due to a poor decision is not clear. What is clear is the amount of horrific refereeing decisions made during the tournament, ranging from inconsistent awarding of cards, poor offside decisions and the infamous Frank Lampard disallowed goal against Germany. However, maybe it was a good thing, as it provided irrefutable evidence to Sepp Blatter that his “technology in football is evil” stance was lunacy. Well done Sepp for working out what everyone else in the world already knew.

Is 9.875 corners per game statistically different from 11?

Since I’ve been in a statistical mood recently, I thought I’d formally test whether my hypothesis of 11 corners per game was a ‘good’ one. (I’m not going to explain the mechanics behind what I am doing, as it is a complex topic that requires reasonably detailed study. The conclusion is the most important thing for those of you that don’t understand what I’m doing!)

  • Null Hypothesis: Number of corners per match = 11
  • Alternative Hypothesis: Number of corners per match is not 11

My sample size is 64, since that’s how many matches were played in the World Cup. Therefore the degrees of freedom under the t distribution is 64 – 1 = 63.

The standard error of the sample is 3.516. (This is an estimate of the standard deviation).

The t-statistic is given by (9.875 – 11) / 3.516 = -0.32.

The negative critical value for t(70) at 5% significance level is -1.99. This is lower than the critical value for t(63) and therefore we can (comfortably) fail to reject the null hypothesis that the number of corners per match is 11, since our test statistic is not below this value.

Hence, I can say that my prediction of 11 corners per match holds pretty firm. 9.875 corners per game is statistically close enough to 11 not to disprove my theory. So overall, my predictions were not too shabby!

Leave a comment

Time limit is exhausted. Please reload CAPTCHA.