Recently in Forecasting Category

You don't say...

What do you mean, municipal wi-fi networks aren't the unalloyed good they were sold as?

Across the United States, many cities are finding their Wi-Fi projects costing more and drawing less interest than expected, leading to worries that a number will fail, resulting in millions of dollars in wasted tax dollars or grants when there had been roads to build and crime to fight.

I was recently in Pittsburgh, PA. Which has a downtown network that can be used "free" for two hours. I say "free", since I had to register. For the benefit of sending my info to the city, which I assume logged my IP and thus knew roughly where I was and more...I got slow service that could only be used in certain positions since the repeaters were weak and stationed poorly for coverage.

But, you know, it's hard to tell how these things will pan out.

Wondering who's gonna win this year's Superbowl? Just ask eBay.

The company Mpire is plotting the price of team paraphenalia for the Seattle Seahawks and the Pittsburgh Steelers, in an attempt to see if the price might be a decent predictor of Superbowl victory. The intuition, I suspect, is that people will be buying the merchandise of the team they expect to win, in the hopes that it will increase in value once the win occurs. Aggregating the data into a single price is intended to look something like the confidence that the crowd has in a win by one team or the other.

From a story on the experiment:

The confounding success of the famous Super Bowl economic indicator, where the winning team's conference seems to correlate with the coming year's economy, is spilling over into other areas of voodoo economics. This time it's reversed as eBayers are predicting the outcome of America's time-honored obsession based upon the going price of each team's merchandise.

Right now, Seattle is around $150, while Pittsburgh hovers near $100. Since Pittsburgh is generally viewed as the favorite, I have to wonder if there aren't a lot of people spending rather small amounts of money on Seahawks trinkets just on the off chance that an upset happens. After all, depending on the consumer's belief in the increase of the value, the expected value of the Seahawk do-dad could be above that of the Steeler gee-gaw, no matter what team the consumer believes will win. (Hey, wait -- does that mean we should read this prediction in reverse? Since the crowd seems to favor Pittsburgh, we should expect Seattle-stuff to sell for more? Not sure. My guess at relevent parameters: 1) closeness of the line, since a wide spread might induce people to forget the long-shot bet and snap up the favored-team's stuff before everyone else does, and 2) the price of the item purchased, since a signed t-shirt might go for cheap now, but could increase several times over in value, whereas expensive items require a much greater level of certainty. Anyway...)

Brain Damage = Better Forecasting?


HedgeFundGuy has an interesting post up about why experts don't often exhibit extraordinary skill at forecasting.

All well and good, but in an appeal to Occam's Razor, perhaps this simpler explanation is the better one?

A team of U.S. scientists has found the emotionally impaired are more willing to gamble for high stakes and that people with brain damage may make good financial decisions, the Times newspaper reported Monday.

In a study of investors' behavior 41 people with normal IQs were asked to play a simple investment game. Fifteen of the group had suffered lesions on the areas of the brain that affect emotions.

The result was those with brain damage outperformed those without.

Perhaps if people can be rational in the grips of a drug addiction -- to note the work of a newly-minted Genius* (almost unrecognizable in that picture, give the lack of what I thought to be the ever-present ballcap) -- then they can be just as rational in the face of brain damage. Indeed, it may be a common occurance:

Fellow [study] author, Baba Shiv of Stanford Graduate School of Business said many company chiefs and top lawyers may also show they share the same trait.

* I note that MR has a post on Kevin Murphy as well, with appropriate links. Tyler Cowen is right to mention the appellation "genius" is deserved even without John D. and Catherine T. say-so. Attending the University of Chicago without, at the very least, sitting in on some of his classes would be a downright shame. Even more rewarding are those days when someone in the audience takes it in their mind to challenge Dr. Murphy on a topic he or she is just learning. The resulting exchange is entertaining, though one must enjoy a bit of shadenfreude for the unfortunate student.

Starbucks Messes With Traffic Models


Another attempt to predict a complex system -- foiled by Starbucks and Dunkin Donuts:

In addition to its effect on mileage, the boom in on-the-go breakfasts has confounded attempts to forecast travel patterns, which are based on computer models that rely heavily on the predictability of the morning commute. Those models assume that people take the shortest, fastest routes to work, not the ones that necessarily lead past a doughnut shop.

"How do we predict future travel when commercial and social interactions like this can surprise us?" McGuckin asked.


Powered by Movable Type 5.02

About this Archive

This page is an archive of recent entries in the Forecasting category.

Food is the previous category.

Gambling is the next category.

Find recent content on the main index or look in the archives to find all content.