World Cup results: Uncertainty of outcome confirmed but....

Dr Nicolas Scelles

Posted: July 3, 2014

In my previous postings, I suggested 2 kinds of predictions. At the end of the first round, what about my first predictions? Out of the 48 games, I have 6 exact scores, what means that in betting on this basis, I would have made a profit. My predictions were 3-0 or 3-1 (+2.51) for Brazil-Croatia (3-1) and 0-1 or 0-0 (-0.51) for Japan-Greece (0-0) so I could have 8 exact scores. If we focus on differences and not exact scores, I have 12 good predictions – almost 14 with the 2 games above. For 17 games, I made an error for 1 goal. It means that in 2 cases out of 3, I was right or at 1 goal of the good difference. But 12 good differences also mean that I was wrong in 3 cases out of 4! And 6 exact scores mean 7 bad scores out of 8! Uncertainty of outcome confirmed!

During the first round, the percentage of game-time with a difference between teams of no more than 1 goal (intra-match balance) was 86.0%, consistent with previous data for the World Cup and uncertainty of outcome. The hierarchy among the 8 groups for intra-match balance is as follows: 1. D (Costa Rica, England, Italy, Uruguay) 98.9%, 2. C (Colombia, Greece, Ivory Coast, Japan) 92.0%, 3. F (Argentina, Bosnia and Herzegovina, Iran, Nigeria) 90.9%, 4. G (Germany, Ghana, Portugal, USA) 89.3%, 5. H (Algeria, Belgium, Russia, South Korea) 88.5%, 6. A (Brazil, Cameroon, Croatia, Mexico) 81.9%, 7. B (Australia, Chile, Netherlands, Spain) 78.9%, 8. E (Ecuador, France, Honduras, Switzerland) 68.0%. Interestingly, the hierarchy for balance based on final standings (standard deviation of the number of points for which the optimal value is 0, meaning that all the teams are equal) is not the same than for intra-match balance: 1. G 2.12, 2. D 2.38, 3. E 2.68, 4. A, C and F 2.95, 7. H 3.08, 8. B 3.35 (the worst value: Netherlands beat Chile, Spain and Australia, Chile beat Spain and Australia, Spain beat Australia = larger disequilibrium between teams). 1 certainty: group B, presented before the competition as perhaps the most difficult with the last two finalists, was eventually not so balanced.

Besides, the percentage of game-time with “real” uncertainty was only 77.0%. Indeed, no uncertainty in Australia-Spain (two teams already eliminated), England-Costa Rica (England already eliminated and Costa Rica never at one goal of losing its first rank) and only 4 minutes with uncertainty in Bosnia and Herzegovina-Iran (with 1-1 in Argentina-Nigeria after 4 minutes, Iran needed to score two goals and not one anymore to be second if Nigeria conceded a goal – Bosnia and Herzegovina was already eliminated; when Nigeria conceded its second then third goal, Iran was led so one goal for Iran and another goal conceded by Nigeria would have not been sufficient for Iran). The hierarchy among the eight groups for “real” uncertainty is as follows: 1. A 85.7%, 2. D 82.2%, 3. C 81.9%, 4. H 78.9%, 5. G 74.6%, 6. F 72.6%, 7. E 71.9%, 8. B 63.7%. Note that groups A and E have a better “real” uncertainty than intra-match balance: even with a difference of more than 1 goal, some games were uncertain (Brazil-Cameroon and Mexico-Croatia for the first rank, Switzerland-Honduras because a goal by Ecuador against France could have eliminated Switzerland even when Switzerland led 3-0 against Honduras if Honduras would have scored).

When “real” uncertainty is too far from intra-match balance, it means that the format of competition is not optimal since it does not prevent games without stakes and/or time without “real” uncertainty. A non-optimal format means that the percentage of prizes can be increased. For the World Cup, it suggests a larger number of teams qualified for the next round. Consider the case where 24 teams are qualified instead of 16 (75%). Spain-Australia would have been with a prize. England would have hoped to be third in its group. Bosnia and Herzegovina-Iran would have not been without “real” uncertainty after 4 minutes. But is it fair to qualify teams that are not in the first half of their group? What about the organisation of an additional knock-out round (second against third before the round of 16)? Could additional dates be found? Would the richest clubs accept this whereas some of their players could be in contention for a larger duration during the World Cup?

In my previous research, I suggested that the qualification of two teams out of three (67%) is perhaps a never-to-be-exceeded cap. It was the case in the World Cups 1986, 1990 and 1994 with 16 teams out of 24 qualified for the round of 16. It will be the case in the Euro 2016. For the World Cup, we could have 20 teams qualified out of 32 (62.5%) with the 12 best teams directly qualified for the round of 16 (playoffs between the last four second teams and the first four best teams). Other possibilities could be related to an increase in the number of teams, 24 teams qualified out of 36 (67%) or 40 (60%) for example. This aspect is not neutral not only for the “real” uncertainty of the World Cup but also for the election of the future FIFA president in 2015: Sepp Blatter would like more African and Asian teams at the expense of European and South American teams, Michel Platini proposed that the World Cup finals should be expanded from 32 to 40 teams. Uncertainty of outcome and format of competition could be important stakes in future debates!



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.