We've all learned from TV and movies that when a serial killer is on the loose, an attractive outside expert can come in and discover an intimate window into the killer's mind by examining the very pattern of his knife strokes.
"The body was found outside, which means our killer can't possibly be a white man."
How does a profiler pull off this magic? According to some studies, they actually don't. After analyzing studies on criminal profiling accuracy, the authors concluded that professional profilers don't show any more significant accuracy in their predictions than the control groups did by using common sense and educated guesses. Also, many profilers refuse to participate in any kind of study to verify their accuracy.
"A study, you say? Well fuck your tits, madam, I have a book to write."
Elusive as they are to study, it's hard to say for sure how good criminal profilers are. Some have certainly been less successful than others, like the FBI profilers hunting the Unabomber, who identified their suspect as a married man living in a house in the suburbs, most likely an airplane mechanic. He was finally arrested in 1996 at his remote cabin where he had been living as a wild-haired, crazy, mountain man for 25 years.
Not pictured: a wife, an airplane.
Many self-proclaimed criminal profiling experts also shoved their faces into the media spotlight during the Washington Beltway sniper attacks to peg the randomly murdering snipers as a couple of white guys. "The experts were neither misogynists nor racists. They all agreed with Van Zandt that 'this is something white males do.'"
They slipped back into the shadows when John Allen Muhammad and Lee Boyd Malvo, two black men, were arrested and ultimately convicted for the killings.
While it's long been a running joke that TV weather forecasters are hired for their good looks or entertainment value, we assume that someone in the back room is feeding them accurate information so they can at least read the weather.
A bored and curious gentleman in Kansas City with a penchant for statistical analysis decided to explore this assumption one day. He tracked the predictions of four local stations over 220 days and found that the four stations had about an 85 percent success rate in predicting if it would rain the next day, which looks pretty good at first glance.
But, here's the thing: It doesn't rain on most days. It's not a 50/50 thing. In most parts of the country it only rains on about 14 percent of the days.
OK, so suppose you went on the air and just predicted every single day that it won't rain. In this 220 day study the four TV stations went through, you'd beat their average accuracy rate (since your "it won't rain" prediction is right 86.3 percent of the time).
Two out of the four stations barely beat you (they got 87 percent) while the other two fell below the threshold.
The study then narrowed it down to only debatable days, eliminating days when it clearly wasn't going to rain--basically boiling it down to days people would actually care about the forecast. He lowered his threshold to 50 percent, the equivalent of flipping a coin when it's cloudy to predict whether it will rain tomorrow, and the news stations again barely managed to defeat the inanimate object, ranging between 50 percent and 60 percent accuracy.
Seriously, anyone can do this.
By the time the test was adjusted to predict the weather three days out, the coin was winning in all cases.
With all that said, this guy's conclusion isn't that meteorology is untrustworthy, but rather that local TV weather forecasting places too much emphasis on good hair and terrible jokes and not enough on smaller details, such as when it is going to rain.
Millions of guys would love to spend all their time watching games and telling people their opinions about sports, but only a select few get to do it, and they do so partly by keeping up a pretense of having some exclusive knowledge about the game that no one else does.
Any sports fan will tell you what a retarded hack their hometown sports columnist is, but sports fans (as with fans of anything, really) tend to be just as lazy as they are abusive, and not many compile a statistical analysis of their hated sportswriters' inaccuracies.
One man, however, did take it upon himself to prove the point empirically in 1971 with an actual study on sportswriters' ability to predict college and NFL games. Their success rate was .476, which you may notice is slightly worse than a coin. The coin's writing ability is arguably superior.
Before writing sports journalists off as complete morons, keep in mind that even Accuscore, a service that charges for its sports predictions based on complex computer algorithms that crunch stats and predict trends, only claims about 53 to 54 percent accuracy, which is still enough to make its customers money.
So, sports prediction is something that almost nobody can get a handle on, but still... worse than a coin toss?
If you want to tie your brain in a knot, think about this: If those guys sitting behind the desk at ESPN are performing worse than chance when they try to make an "expert" judgment about who's going to win the game, that means they could improve their accuracy by always betting on the team they actually think is going to lose. Hell, some of them are wrong so often they could beat the Accuscore service simply by going against their instincts every time.
Eh, they'd still probably fuck it up somehow.
Do you have something funny to say about a random topic? You could be on the front page of Cracked.com tomorrow. Go here and find out how to create a Topic Page.
And check out why you should never trust advertisements, in The 10 Most Laughably Misleading Ads and The 5 Creepiest Advertising Techniques of the (Near) Future.
And stop by our Top Picks (Updated 1.26.2010) brought to you by an "Internet Expert" (read: guy who looks at porn all day).