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?"

Comments

We have to!

[DWMorse]

We just bailed out the banks, so it's too late to start throwing in votes of no confidence!

Re:We have to!

[SimonTheSoundMan]

I much prefer the Institute of Engineering and Technology "The science of predicting football results".

Re:We have to!

[JackieBrown]

(Also, very few of the banks were actually "bailed out." Most were provided with a government-backed high-interest loan that they had to pay back, which many already have)

And not all of the banks wanted the loans but the government pushed it on more banks than were necessary so that the consumers would not know which banks were in need of a "bail-out." The government was worried that if we knew which banks were in need of help, we would all leave those badly run banks in favor of the banks that handled our money responsibly.

In Texas, that is actually a selling point in a number of local bank commercials "And you can rest assured that we did not accepted any government bail out money. Our customer's investments were always secure."

Re:We have to!

[apoc.famine]

Everyone. Every mistake, every poor choice. Just like "The Hand of God", and Germany's "notched goal line", the drunken trader who improperly traded seven million barrels of oil while hammered out of his gourd is reality.

Predict all you want - you can have the best algorithms and the best data, but human fuckups will ruin those every time. Maybe, if all trades were designed and executed by computers, with no human involvement, you could predict those things. But as long as someone is required to spot someone 3 yards offside, or someone is allowed to trade millions of shares with his right hand with a scotch in his left, you're going to be unable to predict either a World Cup or a financial market.

Humans are ridiculously chaotic and unreliable, despite how advanced we think we are.

Re:We have to!

[ducomputergeek]

My best friend in college came from a family of bankers. They own 2 banks. The bigger of the two is still family run, but he owns more of the smaller bank, which is more of a rural community bank. Well apparently, both banks were called by the local fed branch and told to take Tarp money. The bigger bank did. The smaller bank said, "No thank you, we didn't make bad loans." Well the smaller bank has been dragged through the mud with 5 separate investigations in the last 18 months by the government. They'll hold local press conferences saying, "This bank is under investigation for violation of the Equal housing laws, or this or that". Each time, the investigations have come up empty or they've been cleared of wrong doing. But is there any press conferences about that from the feds? Nope. They didn't do anything wrong other than refuse to take the government's money and ran their bank wisely.

Apparently, if you didn't need Tarp money, the government look at your bank as though you MUST have been doing something "wrong".

Re:We have to!

[ultranova]

They got bailed out to avert a total economic collapse -- not as a vote of confidence. If anything, the bailout was a symbol that the banks fucked up -- badly.

This rises a question of whether banks should be simply nationalized? They aren't acting as private companies anyway, since the risk is public, and we already paid for them. Taking them under direct and permanent governmental control would allow us the people to exert some control over them and prevent this from reoccurring.

No private company should be allowed to be so large that its owner can blackmail the rest of the nation. I, for one, do not think that billionaires have any more divine right to rule than kings.

Re:We have to!

[im_thatoneguy]

This is what drove me crazy about watching Star Trek TOS. Spock would always try to cover is his irrational and mentally handicapped shortcomings by blaming it on 'irrationality'. The worst was when Kirk beats him at chess. "That move was illogical." Yo. Spock. If he beat you, it was obviously the most rational move.

To not take into consideration the motivations and irrationalities of the agents in a system is to not model the system at all.

There are a lot of straw Vulcans in the financial market. They reject the idea that people don't always look out for their own best interests or have insufficient education to actually make an informed decision.

hmm amaterism

[cruonit]

the problem is they didn't used enought feature factors or the relevant ones (like detailed statgistics for every player, for that player with coop with another player). i didn't read the stuff but i am sure they used only few factors mostly historic data the detailed forcast would cost too much money (but they invest them into stock forecast)

Best prediction so far:

[Anonymous Coward]

That squid or octopus or kraken Paul which consistently picked the "German" mussel.

What about me ?

[zero.kalvin]

I predicted 3 of the 4 teams in the semi-finals. And I am not claiming any special powers here. But this does raise two issues. If these banks actually were that serious with the prediction: Either they are insane to actually try to predict the outcome using models and techniques that are used for finance, which raises the question of how fit they are to think they can predict the markets themselves if they think they have a "pass-partous" models. And if did they take into account every variable in the prediction, with such a poor predictions I can safely say they cannot be trusted with the financial future of a kid's lemonade stand.

Re:What about me ?

[Anonymous Coward]

You're being silly here.

You're saying because they used financial methods to figure out something that isn't financial, it makes them or their models something we cannot trust.

This is like saying, "Oh daaaam, the chemistry phd tried using chemistry to make himself a meal. The meal tasted like *shit* so his chemistry must be terrible too!"

The problem here isn't that the banks are bad at modeling. The problem is that sports, and anything requiring *small* numbers of people, is hard to predict due to individual differences, wind, vuvuzela interference, etc. If each team had instead had a million people each, statistical modeling would have been far more accurate.

prediction?

[chichilalescu]

I thought the point of the stock market was that people with money can buy shares into companies they think will be profitable. i pitty these idiots who try all their best to get as rich as possible as fast as possible. and I pitty the rest of the world, who see clearly that these guys treat everything like a lottery, and still trust them.

The world is still interesting

[RichMan]

The predictions by various teams would have had a chance of happening.
Like Argentina will beat Germany 67% of the time. There is still the possibility that Germany will win a game.
One thing that makes the world cup very unpredictable is that only a single game decides winner/loser. Anything can happen in a single game. Someone can be a little off. The ball can bounce just so. A ref can blow a call. If they played best of 5 or 7 games the predictions would have a better chance of happening.

Still we can't predict the future 100% and the world is still interesting.

To the stock market. There are to many outside influences.
Could anyone have predicted the well blow out in the Gulf and its affect on the Gulf fishing industry or 911 ? Sure they would have been outside possibilities but no one could have predicted exactly when based on just looking at the stock market.
As much as the economists like to assume that economics is a measurable science the ideas of perfect knowledge and perfect actors are laughable given the way we know people operate. The basic foundation of the economic theory is very broken.

Footy tipping

[MichaelSmith]

...lots of people do it in Australia, especially in work places. Maybe you get a small prize if you guess best. Anyway one year we had this French manager who won the competition by a mile. All he had was a table of match results and a copy of excel. He was also the software metrics guru so knowing how to drive the spreadsheet helped as well.

I think the secret of his success was what he left out of his model. It wasn't smart at all. Just that when teams A and B play, what is the probability that A will win, or something simple like that.

Re:Upgrade

[Anonymous Coward]

use this design

Re:No.

[causality]

So stop. And stop voting for the people that do as well.

There is no chance of that happening in the USA as long as the two major parties are the gatekeepers of federal elections. The "lesser of two evils" is still evil, and election after election of some kind of evil adds up to a lot of institutionalized evil.

Re:No.

[mobby_6kl]

"No" is not the answer, Paul is. He correctly predicted the outcome of all of Germany's games so far, whichi is pretty impressive, considering Germany's loss to Serbia.

Of course, Paul is an octopus who picks the winning side by choosing the box with the winning side's flag on it.

Re:Look at the numbers first

[Kjella]

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

It's been clear to me now for some time that soccer isn't about making the best team win. If a typical match result was like 9-4 it's very unlikely the weaker team would beat the stronger, and we could have done that by making scoring easier. But with results like 2-1 it's pretty much down to $random circumstance of the day. Why? Because in 80% of the matches it makes fans go "If only..." then our team would have won. You look at the one missed chance and ignore that the opponent had three. At 13% you realize this is mostly luck, it means there's at least ten countries that could win.

Personally I think Germany will pull it off though, impressive play and all their hardest competitors eliminated. Spain has Barcelona which is possibly the world's best team but their national team never quite made it, Netherlands and Uruguay have always been good nations but haven't made it to the top in ages. Because when you look at the list of winners, it's shorter than you might think and Germany is the only one left that has won in recent history (Uruguay won in 1930 and 1950).

They use advanced statistical analysis

[Anonymous Coward]

The banks hire actuaries by the boatload to create advanced statistical models which are then programmed in the most efficient manner in order to predict the outcome of any given event. They present the data to the sales team to sell to customers over the phone. Billions are traded every day in this manner. The only possible problem is..... that their models may not actually reflect the real world, and worse..... someone comes along and breaks the perfect statistical distribution scenario ... like "too big to fail" or "labor strike" or "stimulus package" or any of a thousand other events that breaks the standardized distribution model and pisses all over the stock ticker parade. "Bank bailout", "labor shortage", "volcano eruption", .... pick another, there are thousands to chose from. Every last one will work on the market, and they all work in concert. Oh, and as stated before, sometimes the banks just get it wrong..... the question that you really have to ask is: do they every get away from 50:50?

Re:Look at the numbers first

[PietjeJantje]

Netherlands and Uruguay have always been good nations but haven't made it to the top in ages.

You really should shut up if you know nothing about football, because the Netherlands is the most consistent top football country. If you say what you said you are a total non-expert. I think the USA will not win the next Basketball gold, because they haven't made it to the top in ages. Similar statement.

Re:No.

[Anonymous Coward]

Agreed. Take a look at their current predictions. Take a look at how the boom they built turned out.

They're selling 21st century snake oil, and investors are buying it by the barrel. Investment firms (read as "gambling corporations") should NOT be part of boring old banks. Ban combinations of the two, and shoo the investment firms away from the Fed's cheap money window forever.

Re:We can't trust them becasue econ isn't science.

[RyuuzakiTetsuya]

Economics is applied sociology. Like political science. It's just a science where the variables are largely unknown. Which males models and predictability difficult.