Tomorrow marks the start of the World Cup with Brazil playing Croatia in the first game of the competition. Over the next month, 32 teams will compete for the FIFA World Cup trophy. Spain are the favorites according to the FIFA rankings.
Like the stock market day to day, short term events like a 90-minute soccer match can be hard to predict. The fact that that the World Cup is primarily a knock-out competition after the group stage doesn’t help, and the prospect of penalties further complicating the picture.
However, as with the stock market, looking over the longer term, World Cup predictions become clearer. It is easier to predict the outcome of a tournament than the outcome of a single game. It’s also easier to predict the aggregate outcome over multiple World Cups than a single winner.
Of the next several world cups it’s a fair bet that countries with a strong soccer tradition and heritage of past success such as Brazil, Italy, Germany and Argentina will do well. These countries have been at least one of the finalists in 18 of the last 19 World Cups and so are likely to have a strong showing in future years. In addition, being host clearly helps, with the host country winning the tournament just under a third of the time (though of course, having a strong soccer tradition does mean you’re more likely to host the tournament in the first place).
However, climate does matter, the winner comes from the continent the tournament is played in almost 80% of the time. These statistics bode well for Brazil and Argentina who are both in the top 5 of the FIFA rankings currently with the tournament being hosted in Brazil. It’s also worth noting that the finalists are somewhat easier to predict than the semi-finalists, which have contained lots of relative wild cards such as Bulgaria, Sweden, Turkey and South Korea in the last decade. This is because the knock out stage lasts only 4 games, and so it takes only 3 wins to reach the semi-finals based on probability models such as the binomial distribution the chance of an unexpected team reaching the semi-finals is about three times as great as that same team making the final.
Similar logic applies to the stock market, day to day changes are basically impossible to predict, over the course of a year you may begin to have a chance, and over a decade forecasts become potentially more reliable. For example the Nobel prize winning economist Robert Schiller wrote a paper showing how the P/E ratio of the market can be a useful forecast of returns over a decade or more. Over shorter time periods, economists have had much less success with their predictions.
It is a little counter-intuitive, but in many fields predictions that span a longer period of time have a greater accuracy than shorter term predictions. The reason is that the noise of short term ‘random’ events tends to be less prominent over the longer term.