Prediction

Prediction

The science (or art) of predicting 2014 World Cup scores

Dr Nicolas Scelles

Posted: June 23, 2014

In my second posting, I provided my predictions for the World CupTM based on scores over the period 08/2012-12/2013 thanks to 1 regression including intra and inter-confederations games without distinction. Nevertheless, we can think that teams have not the same level of performance as a game is intra-confederation or inter-confederations. Consequently, I could have based my predictions on 2 regressions: 1 for intra-confederation games and 1 for inter-confederations games. With these 2 regressions, here are my other predictions:

 

Match Teams Score
A1 Brazil-Croatia 0 (-0.08)
A2 Mexico-Cameroon +2
A3 Brazil-Mexico +2
A4 Cameroon-Croatia -4
A5 Cameroon-Brazil -4
A6 Croatia-Mexico +2
A: 1. Croatia 7 (+6), 2. Brazil 7 (+6), 3. Mexico 3 (-2), 4. Cameroon 0 (-10)
B1 Spain-Netherlands 0
B2 Chile-Australia +4
B3 Australia-Netherlands -2
B4 Spain-Chile 0
B5 Australia-Spain -4
B6 Netherlands-Chile -1 (-1.498)
B: 1. Chile 7 (+5), 2. Spain 5 (+4), 3. Netherlands 4 (+1), 4. Australia 0 (-10)
C1 Colombia-Greece +2
C2 Côte d’Ivoire-Japan 0 (+0.26)
C3 Colombia-Côte d’Ivoire +2 (+1.504)
C4 Japan-Greece 0 (+0.17)
C5 Japan-Colombia -2
C6 Greece-Côte d’Ivoire 0 (-0.43)
C: 1. Colombia 9 (+6), 2. Côte d’Ivoire 2 (-2), 3. Japan 2 (-2), 4. Greece 2 (-2)
D1 Uruguay-Costa Rica +3
D2 England-Italy +1
D3 Uruguay-England 0 (-0.17)
D4 Italy-Costa Rica +2
D5 Italy-Uruguay -1
D6 Costa Rica-England -3
D: 1. England 7 (+4), 2. Uruguay 7 (+4), 3. Italy 3 (0), 4. Costa Rica 0 (-8)
E1 Switzerland-Ecuador +1
E2 France-Honduras +1 (+1.493)
E3 Switzerland-France -1
E4 Honduras-Ecuador 0
E5 Honduras-Switzerland -2 (-1.508)
E6 Ecuador-France -1
E: 1. France 9 (+3), 2. Switzerland 6 (+2), 3. Ecuador 1 (-2), 4. Honduras 1 (-3)
F1 Argentina-Bosnia and Herzegovina +3
F2 Iran-Nigeria -2
F3 Argentina-Iran +3
F4 Nigeria-Bosnia and Herzegovina +2
F5 Nigeria-Argentina -1
F6 Bosnia and Herzegovina-Iran 0 (+0.08)
F: 1. Argentina 9 (+7), 2. Nigeria6 (+3), 3. Bosnia and Herzegovina 1 (-5), 4. Iran 1 (-5)
G1 Germany-Portugal +1
G2 Ghana-USA -1
G3 Germany-Ghana +1
G4 USA-Portugal 0
G5 USA-Germany 0
G6 Portugal-Ghana +1
G: 1. Germany 7 (+2), 2. USA 5 (+1), 3. Portugal 4 (0), 4. Ghana 0 (-3)
H1 Belgium-Algeria +3
H2 Russia-South Korea 0
H3 Belgium-Russia 0
H4 South Korea-Algeria +4
H5 South Korea-Belgium +1
H6 Algeria-Russia -4
H: 1. South Korea 7 (+5), 2. Russia 5 (+4), 3. Belgium 4 (+2), 4. Algeria 0 (-11)
M49 Croatia-Spain -1
M50 Colombia-Uruguay 0 (+0.09)
M51 Chile-Brazil -1
M52 England-Côte d’Ivoire +2
M53 France-Nigeria +1
M54 Germany-Russia +1
M55 Argentina-Switzerland 0 (-0.01)
M56 South Korea-USA +1
M57 France-Germany -1
M58 Spain-Colombia 0 (+0.425)
M59 Switzerland-South Korea +0.5
M60 Brazil-England +1
M61 Spain-Germany -1
M62 Brazil-Switzerland +2
M63 Spain-Switzerland +1
M64 Brazil-Germany +4

 

Compared to my previous predictions, a lot of changes can be noticed: Germany in final and Switzerland in semi-finals instead of Argentina; France, Switzerland, South Korea and England in quarter-finals instead of Ecuador, Argentina, Russia and Uruguay; England, Nigeria, Switzerland, South Korea and USA in the round of 16 instead of Italy, Bosnia and Herzegovina, Ecuador, Belgium and Portugal. Nevertheless, at the end, Brazil should still be the winner with an overwhelming advantage of 4 goals against Germany due to a supposed low level for this latter in inter-confederations games. I write “supposed” because a first weakness of the approach described here is that inter-confederations games included in the regression are essentially friendly games. It is not sure that Germany (and also other teams) played at its (their) best level in these games.

Another weakness is the lack of inter-confederations games for the main part of teams: 1 game for Iran (2-2 vs Tunisia in Hungary), 1 game for Algeria (a defeat 1-0 at home vs Bosnia and Herzegovina which lost its 2 other inter-confederations games – 4-3 vs USA at home and 2-0 vs Argentina in USA; it explains the very bad coefficient for Algeria), 2 games for Croatia (2 wins vs South Korea – 4-0 at home and 2-1 away – which won 2-1 at home vs Switzerland which won 1 of its 3 inter-confederations games 1-0 at home vs Brazil; it explains the very good coefficient for Croatia), no game for Greece (coefficient for 2011-2013 instead of 08/2012-12/2013 with only 2 games) and 13 teams with only 3 or 4 games. Of course, the prediction for Cameroon-Croatia (and for Switzerland-Ecuador and Russia-South Korea) is better here (-4 was the actual difference) than previously (-1) but it is not possible to ensure that Croatia is the best team in inter-confederations games (its defeat 3-1 vs Brazil seems indicate that it is not the case, even in taking into account home advantage for Brazil – I do not discuss about refereeing here!). The 2 main weaknesses noted above explain why I favoured a global approach in my previous posting rather than the approach described here.

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.