Prediction

Prediction

Can we explain World Cup outcomes?

Dr Nicolas Scelles

Posted: July 2, 2014

My previous postings were about predictions and uncertainty. In this one, my aim is to envisage whether we can explain World Cup outcomes. In other words, can we find a logic in outcomes for every team or are some teams’ outcomes out of logic? As for my predictions, my approach is based on regressions. I did four regressions: based on scores in 2011, 2012, 2013 and 08/2012-2013. These regressions allowed me to have coefficients for every team and every period. These coefficients correspond to their relative strengths. As the averages for the 32 teams taking part in the World Cup were not the same according to the period, I corrected coefficients for 2012, 2013 and 08/2012-2013 so that their averages were the same than 2011. My aim was to observe how every team’s strengths have evolved during the last three years and how it can explain their World Cup outcomes. Contrary to my previous predictions based on the same period for every team, my idea here is that all the teams have not the same evolution: some could be regular, regularly increase or decrease their level, have cycles. The table below presents the coefficients for the four regressions (that you can consider as the team level on a scale of 0 to 6) and my own estimation for the World Cup based on actual outcomes and my subjective perception (note that home advantage – around +0.5 goal in my four regressions – is not added for Brazil which is all the same the best team).

 

  2011 2012 2013 08/2012-2013 2014 WC
Australia 3.430 1.871 1.649 2.023 2.023
Iran 2.132 0.926 3.351 2.013 2.013
Japan 3.427 2.631 2.245 2.299 1.859
South Korea 3.092 1.904 1.895 1.698 1.698
Algeria 1.789 3.233 1.767 2.156 3.211
Cameroon 3.105 1.673 1.845 1.920 1.673
Ghana 3.262 3.537 2.235 2.960 2.399
Ivory Coast 3.578 2.371 2.504 2.399 2.960
Nigeria 2.952 2.524 2.378 2.705 2.705
Costa Rica 2.085 2.661 2.031 2.371 2.715
Honduras 2.063 3.003 1.820 2.302 1.820
Mexico 3.511 3.669 2.275 2.608 3.669
USA 2.046 3.035 2.937 2.993 2.993
Argentina 3.257 4.991 4.368 4.487 4.368
Brazil 3.853 5.242 4.661 4.851 4.661
Chile 3.392 3.103 4.473 3.803 3.803
Colombia 3.264 4.881 4.256 4.544 4.544
Ecuador 3.205 3.949 3.647 3.801 3.345
Uruguay 4.000 2.953 3.889 3.451 2.842
Belgium 2.646 2.670 3.447 3.243 4.224
Bosnia 2.210 3.737 3.533 3.754 3.329
Croatia 3.114 2.506 3.154 2.862 2.862
England 3.645 3.208 3.674 3.432 2.801
France 2.982 3.415 3.829 3.638 4.262
Germany 3.910 3.368 4.481 4.161 4.161
Greece 2.596 2.751 2.822 2.906 2.906
Italy 3.380 2.695 2.958 2.933 2.933
Netherlands 4.097 3.180 3.751 3.616 4.322
Portugal 3.477 2.457 3.350 3.018 2.330
Russia 3.057 3.387 3.370 3.309 3.309
Spain 4.467 4.517 4.084 4.147 3.651
Switzerland 2.517 3.486 2.860 3.133 3.133
AVERAGE 3.111 3.111 3.111 3.111 3.110

 

The 32 teams could be classified as follows:

18 teams (56%) would play at their level over the period 08/2012-2013 or 2013 (regular): Australia, Iran, South Korea, Ghana, Ivory Coast, Nigeria, USA, Chile, Colombia, Croatia, Germany, Greece, Italy, Russia and Switzerland for 08/2012-2013, Honduras, Argentina and Brazil for 2013.

7 teams (22%) would have pursued their increase or decrease from 2012 to 2013: Ghana, Belgium, France, Netherlands (increase), Ecuador, Bosnia and Herzegovina and Spain (decrease).

7 teams (22%) would be characterised by a cycle phenomenon: Algeria, Costa Rica and Mexico (increase), Cameroon, Uruguay, England and Portugal (decrease).

When comparing my subjective coefficients (I stress on subjective) with actual outcomes, keep in mind two elements related to uncertainty of outcome: the very small margin between a win by 1 goal, a draw and a loss by 1 goal, and the possibility of accident. For example, according to my coefficients, Ivory Coast and Greece would have more or less the same level: Greece beat Ivory Coast at the last minute; Uruguay should have got a draw against Costa Rica whereas it lost 3-1: I consider this as an accident and if the two teams would meet again, I am not sure that Costa Rica would beat Uruguay for a second time (but perhaps I am wrong!). Other example: Mexico should have won 2-0 instead of 1-0 against Cameroon. If we add one of the two Mexican goals disallowed for offside whereas they were not, we have 2-0 which seems to me more representative of the difference between the two teams (of course, the match would have not been the same if one of the two disallowed goals would have been accepted). Colombia would be the second best team behind Brazil (better than Argentina) but as it should face Brazil in quarter-finals, it should not be in final. At the end, Brazil should still be the winner.

About Dr Nicolas Scelles

Dr Nicolas Scelles is lecturer at the School of Sport, Stirling University, Scotland. He holds a PhD in sports economics from the University of Caen Basse-Normandie, France. He has articles in international journals including Applied Economics, Economics Bulletin, International Journal of Sport Finance and International Journal of Sport Management and Marketing.