Prediction for the final: Brazil will defeat Spain | Wired

[PLAYSPORT) THE FORMULA EXPLAINED WINNER Every match, i vs ji, can be forecast by inputting population (pop), GDP per capita (y) and experience (exp) into a formula to give the expected goal difference (GD) or "edge": GDCj) = 0.137 log (pop(/pop() + 0.145 log (y(D/y) + 0.739 log (exp()/Cexp()) (+ 0.657 home advantage when South Africa plays). Figures in the coloured boxes indicate matches likely to be won by each team in the group stage; the formula rarely predicts a draw. Each group's top two teams progress to the knockout, where the same formula is used; GD is shown by the number in black (eg in the semis, Brazil is expected to score 0.39 more goals on average when they play Germany). SERBIA 家 AZIL BRAZIL The World Cup predicted FINAL 0.05 0.39 A country's GDP is a better indicator of its chances in South Africa than its team's playing formation There are many ways t Cup results - the simplest being to use the Fifa rankings in the style of a very basic game of Top Trumps. But in their book Why England Lose (HarperSport), Simon Kuper and Stefan Szymanski develop a model for forecasting international football results based on mostly economic data. Here they use it to predict this year's World Cup. They found that, in almost three in four games, the outcome of a match can be cor- rectly predicted by a formula based on the ratio between the two nations' populations (bigger nations have a larger talent pool), GDP per head (the higher it is, the better the resources), experience (the more you play, the better you get), plus a constant to represent home advantage for the host nation. A doubling of GDP per head or a doubling of population size are each worth about one-tenth ofagoal in terms of advan- tage. That doesn't sound like much, but consider that Switzerland's GDP per head is 100 times larger than Ghana's. For one side to have twice the experience is worth more than halfagoal, and home advantage is worth nearly two-thirds of a goal. The method correctly predicts the winner 72 per cent of the time (and has been weighted to allow for teams doing better or worse than predicted in the past). The algorithm uses linear regression and has been fitted to all international results forecast World ENGLANDm GOAL DIFFERENCE 019 Q42 0.04 - ITALY *** ALGERIA- 8 SLOVENIA O -PARAGUAY SLOVAKIA NEW ZEALAND GROUP KOREA REPUBLIC 2. BRAZILk**** STAGE nN PORTUGAL ARGENTINA NIGERIA - GREECE O e - KOREA DPR CÔTE D'IVOIRE 0.08 I LO3 0.02 026 026 LOI SPAIN FRANCE HONDURAS FITA RANKING SOUTH AFRICA E GROUP WINS between 1980 and 2001 (using City Univer- sity London's football archive). The result- ing figure is interpreted as the expected goal difference; the win is awarded to the team with a positive score. However, if the figure is less than 0.1 in the knockout stage, the authors examine historical per- formance in penalty shoot-outs to break the deadlock. This summer, they foresee England losing to Germany (0.02) early on. Optimists who ignore them can get about 14/1 odds for an England-Brazil final. KNOCKOUT STAGE KEY World Cup wins SERBIA ETHERLANDS KICO ENMARK CTZERLAND SETHERLANDS 3 MONDURAS [PLAYSPORT) THE FORMULA EXPLAINED WINNER Every match, i vs ji, can be forecast by inputting population (pop), GDP per capita (y) and experience (exp) into a formula to give the expected goal difference (GD) or "edge": GDCj) = 0.137 log (pop(/pop() + 0.145 log (y(D/y) + 0.739 log (exp()/Cexp()) (+ 0.657 home advantage when South Africa plays). Figures in the coloured boxes indicate matches likely to be won by each team in the group stage; the formula rarely predicts a draw. Each group's top two teams progress to the knockout, where the same formula is used; GD is shown by the number in black (eg in the semis, Brazil is expected to score 0.39 more goals on average when they play Germany). SERBIA 家 AZIL BRAZIL The World Cup predicted FINAL 0.05 0.39 A country's GDP is a better indicator of its chances in South Africa than its team's playing formation There are many ways t Cup results - the simplest being to use the Fifa rankings in the style of a very basic game of Top Trumps. But in their book Why England Lose (HarperSport), Simon Kuper and Stefan Szymanski develop a model for forecasting international football results based on mostly economic data. Here they use it to predict this year's World Cup. They found that, in almost three in four games, the outcome of a match can be cor- rectly predicted by a formula based on the ratio between the two nations' populations (bigger nations have a larger talent pool), GDP per head (the higher it is, the better the resources), experience (the more you play, the better you get), plus a constant to represent home advantage for the host nation. A doubling of GDP per head or a doubling of population size are each worth about one-tenth ofagoal in terms of advan- tage. That doesn't sound like much, but consider that Switzerland's GDP per head is 100 times larger than Ghana's. For one side to have twice the experience is worth more than halfagoal, and home advantage is worth nearly two-thirds of a goal. The method correctly predicts the winner 72 per cent of the time (and has been weighted to allow for teams doing better or worse than predicted in the past). The algorithm uses linear regression and has been fitted to all international results forecast World ENGLANDm GOAL DIFFERENCE 019 Q42 0.04 - ITALY *** ALGERIA- 8 SLOVENIA O -PARAGUAY SLOVAKIA NEW ZEALAND GROUP KOREA REPUBLIC 2. BRAZILk**** STAGE nN PORTUGAL ARGENTINA NIGERIA - GREECE O e - KOREA DPR CÔTE D'IVOIRE 0.08 I LO3 0.02 026 026 LOI SPAIN FRANCE HONDURAS FITA RANKING SOUTH AFRICA E GROUP WINS between 1980 and 2001 (using City Univer- sity London's football archive). The result- ing figure is interpreted as the expected goal difference; the win is awarded to the team with a positive score. However, if the figure is less than 0.1 in the knockout stage, the authors examine historical per- formance in penalty shoot-outs to break the deadlock. This summer, they foresee England losing to Germany (0.02) early on. Optimists who ignore them can get about 14/1 odds for an England-Brazil final. KNOCKOUT STAGE KEY World Cup wins SERBIA ETHERLANDS KICO ENMARK CTZERLAND SETHERLANDS 3 MONDURAS [PLAYSPORT) THE FORMULA EXPLAINED WINNER Every match, i vs ji, can be forecast by inputting population (pop), GDP per capita (y) and experience (exp) into a formula to give the expected goal difference (GD) or "edge": GDCj) = 0.137 log (pop(/pop() + 0.145 log (y(D/y) + 0.739 log (exp()/Cexp()) (+ 0.657 home advantage when South Africa plays). Figures in the coloured boxes indicate matches likely to be won by each team in the group stage; the formula rarely predicts a draw. Each group's top two teams progress to the knockout, where the same formula is used; GD is shown by the number in black (eg in the semis, Brazil is expected to score 0.39 more goals on average when they play Germany). SERBIA 家 AZIL BRAZIL The World Cup predicted FINAL 0.05 0.39 A country's GDP is a better indicator of its chances in South Africa than its team's playing formation There are many ways t Cup results - the simplest being to use the Fifa rankings in the style of a very basic game of Top Trumps. But in their book Why England Lose (HarperSport), Simon Kuper and Stefan Szymanski develop a model for forecasting international football results based on mostly economic data. Here they use it to predict this year's World Cup. They found that, in almost three in four games, the outcome of a match can be cor- rectly predicted by a formula based on the ratio between the two nations' populations (bigger nations have a larger talent pool), GDP per head (the higher it is, the better the resources), experience (the more you play, the better you get), plus a constant to represent home advantage for the host nation. A doubling of GDP per head or a doubling of population size are each worth about one-tenth ofagoal in terms of advan- tage. That doesn't sound like much, but consider that Switzerland's GDP per head is 100 times larger than Ghana's. For one side to have twice the experience is worth more than halfagoal, and home advantage is worth nearly two-thirds of a goal. The method correctly predicts the winner 72 per cent of the time (and has been weighted to allow for teams doing better or worse than predicted in the past). The algorithm uses linear regression and has been fitted to all international results forecast World ENGLANDm GOAL DIFFERENCE 019 Q42 0.04 - ITALY *** ALGERIA- 8 SLOVENIA O -PARAGUAY SLOVAKIA NEW ZEALAND GROUP KOREA REPUBLIC 2. BRAZILk**** STAGE nN PORTUGAL ARGENTINA NIGERIA - GREECE O e - KOREA DPR CÔTE D'IVOIRE 0.08 I LO3 0.02 026 026 LOI SPAIN FRANCE HONDURAS FITA RANKING SOUTH AFRICA E GROUP WINS between 1980 and 2001 (using City Univer- sity London's football archive). The result- ing figure is interpreted as the expected goal difference; the win is awarded to the team with a positive score. However, if the figure is less than 0.1 in the knockout stage, the authors examine historical per- formance in penalty shoot-outs to break the deadlock. This summer, they foresee England losing to Germany (0.02) early on. Optimists who ignore them can get about 14/1 odds for an England-Brazil final. KNOCKOUT STAGE KEY World Cup wins SERBIA ETHERLANDS KICO ENMARK CTZERLAND SETHERLANDS 3 MONDURAS [PLAYSPORT) THE FORMULA EXPLAINED WINNER Every match, i vs ji, can be forecast by inputting population (pop), GDP per capita (y) and experience (exp) into a formula to give the expected goal difference (GD) or "edge": GDCj) = 0.137 log (pop(/pop() + 0.145 log (y(D/y) + 0.739 log (exp()/Cexp()) (+ 0.657 home advantage when South Africa plays). Figures in the coloured boxes indicate matches likely to be won by each team in the group stage; the formula rarely predicts a draw. Each group's top two teams progress to the knockout, where the same formula is used; GD is shown by the number in black (eg in the semis, Brazil is expected to score 0.39 more goals on average when they play Germany). SERBIA 家 AZIL BRAZIL The World Cup predicted FINAL 0.05 0.39 A country's GDP is a better indicator of its chances in South Africa than its team's playing formation There are many ways t Cup results - the simplest being to use the Fifa rankings in the style of a very basic game of Top Trumps. But in their book Why England Lose (HarperSport), Simon Kuper and Stefan Szymanski develop a model for forecasting international football results based on mostly economic data. Here they use it to predict this year's World Cup. They found that, in almost three in four games, the outcome of a match can be cor- rectly predicted by a formula based on the ratio between the two nations' populations (bigger nations have a larger talent pool), GDP per head (the higher it is, the better the resources), experience (the more you play, the better you get), plus a constant to represent home advantage for the host nation. A doubling of GDP per head or a doubling of population size are each worth about one-tenth ofagoal in terms of advan- tage. That doesn't sound like much, but consider that Switzerland's GDP per head is 100 times larger than Ghana's. For one side to have twice the experience is worth more than halfagoal, and home advantage is worth nearly two-thirds of a goal. The method correctly predicts the winner 72 per cent of the time (and has been weighted to allow for teams doing better or worse than predicted in the past). The algorithm uses linear regression and has been fitted to all international results forecast World ENGLANDm GOAL DIFFERENCE 019 Q42 0.04 - ITALY *** ALGERIA- 8 SLOVENIA O -PARAGUAY SLOVAKIA NEW ZEALAND GROUP KOREA REPUBLIC 2. BRAZILk**** STAGE nN PORTUGAL ARGENTINA NIGERIA - GREECE O e - KOREA DPR CÔTE D'IVOIRE 0.08 I LO3 0.02 026 026 LOI SPAIN FRANCE HONDURAS FITA RANKING SOUTH AFRICA E GROUP WINS between 1980 and 2001 (using City Univer- sity London's football archive). The result- ing figure is interpreted as the expected goal difference; the win is awarded to the team with a positive score. However, if the figure is less than 0.1 in the knockout stage, the authors examine historical per- formance in penalty shoot-outs to break the deadlock. This summer, they foresee England losing to Germany (0.02) early on. Optimists who ignore them can get about 14/1 odds for an England-Brazil final. KNOCKOUT STAGE KEY World Cup wins SERBIA ETHERLANDS KICO ENMARK CTZERLAND SETHERLANDS 3 MONDURAS [PLAYSPORT) THE FORMULA EXPLAINED WINNER Every match, i vs ji, can be forecast by inputting population (pop), GDP per capita (y) and experience (exp) into a formula to give the expected goal difference (GD) or "edge": GDCj) = 0.137 log (pop(/pop() + 0.145 log (y(D/y) + 0.739 log (exp()/Cexp()) (+ 0.657 home advantage when South Africa plays). Figures in the coloured boxes indicate matches likely to be won by each team in the group stage; the formula rarely predicts a draw. Each group's top two teams progress to the knockout, where the same formula is used; GD is shown by the number in black (eg in the semis, Brazil is expected to score 0.39 more goals on average when they play Germany). SERBIA 家 AZIL BRAZIL The World Cup predicted FINAL 0.05 0.39 A country's GDP is a better indicator of its chances in South Africa than its team's playing formation There are many ways t Cup results - the simplest being to use the Fifa rankings in the style of a very basic game of Top Trumps. But in their book Why England Lose (HarperSport), Simon Kuper and Stefan Szymanski develop a model for forecasting international football results based on mostly economic data. Here they use it to predict this year's World Cup. They found that, in almost three in four games, the outcome of a match can be cor- rectly predicted by a formula based on the ratio between the two nations' populations (bigger nations have a larger talent pool), GDP per head (the higher it is, the better the resources), experience (the more you play, the better you get), plus a constant to represent home advantage for the host nation. A doubling of GDP per head or a doubling of population size are each worth about one-tenth ofagoal in terms of advan- tage. That doesn't sound like much, but consider that Switzerland's GDP per head is 100 times larger than Ghana's. For one side to have twice the experience is worth more than halfagoal, and home advantage is worth nearly two-thirds of a goal. The method correctly predicts the winner 72 per cent of the time (and has been weighted to allow for teams doing better or worse than predicted in the past). The algorithm uses linear regression and has been fitted to all international results forecast World ENGLANDm GOAL DIFFERENCE 019 Q42 0.04 - ITALY *** ALGERIA- 8 SLOVENIA O -PARAGUAY SLOVAKIA NEW ZEALAND GROUP KOREA REPUBLIC 2. BRAZILk**** STAGE nN PORTUGAL ARGENTINA NIGERIA - GREECE O e - KOREA DPR CÔTE D'IVOIRE 0.08 I LO3 0.02 026 026 LOI SPAIN FRANCE HONDURAS FITA RANKING SOUTH AFRICA E GROUP WINS between 1980 and 2001 (using City Univer- sity London's football archive). The result- ing figure is interpreted as the expected goal difference; the win is awarded to the team with a positive score. However, if the figure is less than 0.1 in the knockout stage, the authors examine historical per- formance in penalty shoot-outs to break the deadlock. This summer, they foresee England losing to Germany (0.02) early on. Optimists who ignore them can get about 14/1 odds for an England-Brazil final. KNOCKOUT STAGE KEY World Cup wins SERBIA ETHERLANDS KICO ENMARK CTZERLAND SETHERLANDS 3 MONDURAS