Noise News Service

21:08 Mon 09 Jun 2008
[, , , ]

This morning, via A Tiny Revolution, I came across a wire news story from Bloomberg that has the form of informational content but contains almost no useful information.

I don’t mean that it’s full of “useless” information, or trivia, such as data about train model numbers (or MTG card printing errors, etc.). In fact, while purportedly in a much more useful class of information, “news”, this article provides far less than some catalog of geek-only statistics would.

It would also really, really irritate Nassim Nicholas Taleb, whose words on the area of news came back to me strongly when I read it.

The beginning is quite a giveaway in itself:

Crude oil may fall next week because of rising U.S. fuel inventories and declining demand as consumers react to record prices.

It may fall. It may not. If it does, we’ve come up with some plausible-sounding reasons as to why that happens. But we’re not going to say it will fall, or even that it’s likely to fall. No, just that it might.

If you, the reader, need a news article to tell you that crude oil might or might not fall in price next week, then you have some serious problems. But you also apparently fall right into Bloomberg’s target demographic.

Next it tells us that Bloomberg surveyed 29 “analysts”, 18 of whom (64 percent!) said that prices will decline. Last week, though, 48 percent said that oil futures would increase.

Hmm. Fascinating, yes? Especially that last bit. So, if the price of oil futures next week really matter to you, what information do you now have? That 18 of 29 unnamed analysts think that it’ll drop.

Then we’re treated to the wise words of an analyst who gets named in the article, followed by a paragraph about the recent airline industry woes related to high oil prices. This paragraph, and the one following it concerning current oil prices, appear to be the only things in the article that I can regard as useful information, dealing as they do with actual facts rather than pure speculation.

The final sentence really makes the whole thing, though:

The oil survey has correctly predicted the direction of futures 49 percent of the time since its start in April 2004.

49 percent! The Bloomberg Oil Survey, motto: “We’re worse than a coin flip!”.

Unbelievable. They clearly have no clue whatsoever, and their historical record proves this (although note that if their record were really good so far, it wouldn’t prove that they could be trusted to predict prices in the future, but they don’t even manage to raise that issue).

Given that they’re obviously useless, why write the article? Why go to the trouble of establishing some relationship with these analysts, calling them every Thursday, collating the data, writing the article, and putting it out on the wire, when you’d get the same predictive value regarding oil futures if you flipped a coin? Even if they were poor predictors but their attitude reflected the state of the market’s thinking on the subject, you’d expect them to have better than 49 percent success. The only answer I can think of is that someone at Bloomberg thinks it makes the service look good to have an “oil survey” every week, that it provides some kind of competitive edge as compared to other business news services.

That article isn’t “news”. It’s essentially random speculation from a bunch of people who have apparently garnered success in a field where it’s almost impossible to accurately measure/reward actual competence. These are precisely the people Taleb was warning about when he said that if you’re going to criticize people for believing in religion because it’s irrational, you also have to criticize them for paying any attention whatsoever to market analysts—and that you should never trust advice coming from someone dressed in a suit and tie.

Leave a Reply