Monte Carlo for the World Cup

By:
Published on:

Football fans around the world looking forward to running office sweepstakes, emptying online betting accounts and making sure the grad trainee fills in the wall chart each morning of the World Cup that begins on Thursday should thank their lucky stars the tournament can’t possibly be as dull as Goldman Sachs predicts.

The firm’s strategists and analysts have applied to the beautiful game a stochastic model generating distributions for the likely results of the competition’s 64 matches based on a regression analysis of the entire history of international football.

Goldman talks up the statistical power of the Elo database of all football results over the better-known FIFA/Coca-Cola rankings. Unfortunately, though, its analysis suggests that 33 of the 48 group matches will end in 1-1 draws.


Euromoney shares Goldman’s disregard for a FIFA ranking system that puts England as the 11th-best team in the world, but we can’t help feeling its projections are laughably unlikely. Only six of the first 48 matches in 2010 finished 1-1 and just three ended that way in 2010. If this is the way Goldman models financial market returns, investors should beware.

Perhaps conscious it has got off to a disappointing start, Goldman fills out its 64-page report with fawning interviews with Chelsea’s Brazilian players. Let’s hope Peter Oppenheimer, chief global equity strategist, manages more searching questions to financial policymakers than to David Luiz: "Did you imaging playing for your country when you were a child?" Or, even more toe-curling: "The economy in Brazil has been a little bit weak recently, and some people hope that holding the tournament in Brazil will lift spirits and make things better. Do you think that’s true?" Oscar certainly hopes so.

Goldman tells us that after all those 1-1 draws at the group stage, it uses a Monte Carlo simulation to generate a probability distribution of teams reaching the knock-out stages. Here, too, it manages some very implausible-sounding predictions. Fair enough, giving the home nation a 48.5% chance of winning the tournament seems reasonable. The best price at the bookies is 3/1. But Goldman admits its model also predicted a Brazil win four years ago when Spain defeated the Netherlands in the final.

However, Goldman seems to have gone off the rails in giving England a 5% chance of making the final and a 1.4% chance of lifting the trophy, compared with just 0.6% for the much stronger Belgium. Investors should check all Goldman’s market forecasts for any evidence of similarly misplaced over-optimism.