UBS forecasters pegged Germany to win soccer's World Cup. The team crashed out of the tournament in the group stage, raising question marks over the Swiss bank's forecasting ability.
Zurich-based UBS was first and fastest: back in May, the Swiss bank predicted that Germany would win soccer's World Cup – again. Other banks followed suit: Commerzbank, for example, also predicted Germany (the Frankfurt-based lender might be forgiven for home bias).
Banks like UBS take the World Cup a lot more seriously than other sports fans. UBS enlisted its number-crunchers to devise elaborate computer programs using game data and algorithms in order to forecast the likely winner with the help of statistical methods.
Crashing Out
These are, of course, the same programs that UBS uses to predict your investment opportunities – and which invests your money. Tuesday brought a timely reminder of just what these programs and statistical equations are worth: Germany, the defending world soccer champion, ended its World Cup attempt with a humiliating 0:2 defeat at the hands of South Korea.
No computer could have predicted that Germany wouldn't make it past the group stage: the defending world champion has never crashed out of the tournament in the group stage, meaning there simply is no statistical data for it.
Goldman Flip-Flops
UBS' sporting example shows that predictions maybe accurate some times, but at others they're not. He or she who bets on them would be well-advised to take several such predictions into consideration instead of banking it all on one. Thus, potential losses could be diversified – possibly (we're not sure).
To be sure, the Swiss bank is in good company: Goldman Sachs's machine learning-fueled «Soccerbot 3000» initially flagged a German final. The difference is that the American investment bank revised its stance to forecast and Brazil-England final after the tournament began, with the South American country winning the World Cup this year.
