World Cup Prediction Failures

pdcull writes "We all read on Slashdot about the investment banks using their massive computer power and clever modeling techniques to predict the FIFA World Cup outcome. Now that Goldman Sachs's, UBS's and Danske Bank's favorite, Brazil, has been eliminated, and with JP Morgan's England long gone, the question that begs to be asked is: can we really trust these guys to predict the financial markets any better than they did World Cup?"

Look at the numbers first

[Flavio]

Goldman Sachs gave Brazil (the "favorite") only a 13% chance of winning the world cup.

The fact that Brazil was eliminated is not at odds with the reports.

Re:The world is still interesting

[Myji Humoz]

The foundation of economics is the same as the foundation of statistics. No economist can accurately predict how an individual game can turn out, just like no economist can accurately predict how an individual actor will choose. However, the law of large numbers means that given a sufficiently large population, mass scale behavior can be predicted. A simple example of this in action is that if I flip a coin and call it heads or tails, I'll be wrong roughly 50% of the time. However, if I say that 50% give or take 5% of a thousand coin flips will end up heads or tails, I'd probably be right.

Economics does not rely on perfect knowledge and perfect actors. It relies on modeling the real world as best as possible, and you know what? Models usually include abstract representations, and abstract representations include idealized actions.

With regard to the story, it's laughable to think that financial market predictions are anything like sports predictions. These teams are made up of players who don't play together during regular seasons; they're generally on different teams in Europe. How is it news that analyzing various productivity and growth forecasting statistics will fail to predict who wins a 1-0 soccer game?

Re:We have to!

[malkir]

The owner of BP is part of GS.

The predictions involve confidence

[Anonymous Coward]

Well, I think everyone would agree that no prediction is perfect, and all of them have a certain confidence interval associated with any prediction. For example, I find it very disappointing and deceiving to say that Goldman's favorite is out, because the teams they predicted second most likely and third most likely to win the thing are still in it.

Football worldcup predictions

[Anonymous Coward]

Its very hard to predict about any game especially football. Beacuse just one mistake can change everything and good team can also loose.
Also visit: http://top-online-money.yolasite.com

Re:We have to!

[jonbryce]

Paul the Octopus did a better job of predicting the results than they did.

Re:We have to!

[HungryHobo]

Some people have this idea that bankers and investors have an almost godlike knowledge of the world and the future.
On forums which attract financial types I'll see people dismissing stocks on the basis that "of course if they were really more valuable the traders would have snapped then up".
It's as if they don't realize that the traders are operating on almost no information in some cases. living across the street from a company and being able to count the trucks going in and out will often put you ahead of the traders in terms of real information.

Re:We have to!

[tacarat]

It's not the diabolic talent of the customers to refuse to learn more about their money or who handles it? I'm sure most of their customers are only so because they saw a TV ad or something. It's like asking job applicants "honestly, are you a crook" rather than finding out how to do a proper background check.