Dr. Dobb's Journal September 1998
Awakened by a familiar noise, I stumbled out to the kitchen and came upon a sight that was all too familiar. It was the lower two thirds of my cousin Corbett, protruding from my refrigerator. The upper third emerged a moment later and asked, "Got any marmalade?" Too groggy to form actual words, I pointed to the jar of Rose's lime on the second shelf. As I watched him butter and marmalade a slice of toast -- it was the toast popping up that had awakened me -- I tried to pull speech together.
"What are you doing here, Corbett?"
"I've got an early appointment with a vulture capitalist and I need some fuel for the drive over to Sand Hill Road. Are these the only herbal teas you have?"
"Are you starting another business?" I asked, my faculties slowly rallying to the cause as I pulled up a stool to the counter. "What's this one about?"
"It's all about prediction. You ever predict anything that came true?"
"Well, sure. I mean, I predicted that customers would take control of markets."
"Jeez, that hasn't happened."
"Okay, not yet, but it's starting to. I mean, there's a web site called 'priceline.com' where air travel customers tell the airlines what they'll pay. It's standing the traditional market model on its head."
"Whatever. The point is, neurological research has progressed by leaps and bounds since you were in school."
"Oh, is that the point? 'Cause I was wondering."
The tea kettle whistled and he got it, and seemed to find the decaf chai acceptable. "A guy named George Wolford at Dartmouth did this study where he displayed a point of light repeatedly at either the top or the bottom of a computer screen and asked subjects to predict where it would appear next. It was rigged to show up at the top 80 percent of the time, but the actual sequence was random."
"I'm trying to stay with you, Corbett, but..."
"The Dartmouth undergraduates in the study kept trying to figure out the pattern -- a lot of them thought they had figured it out -- and as a result their predictions averaged about 68 percent correct."
"Barely passing, I'd say."
"Wolford also ran the experiment with rats. They didn't try to figure out the pattern."
"Now, we don't know that."
"We know that after a few trials they just pressed the top button all the time."
"Meaning...?"
"Meaning that they were correct 80 percent of the time, significantly better than human beings."
"Than Dartmouth undergraduates, anyway."
He got up to put more toast in the toaster. "Then there's the Concorde fallacy."
"Does that have anything to do with relationships between cousins?"
"It's the mistake of throwing good money after bad. It was either biologist Richard Dawkins or his student Tamsin Carlisle who came up with the name, based on their view that that's what the builders of the Concorde supersonic transport did. Technically, it's the error of basing a decision on prior investment rather than on future prospects. Research shows that human beings routinely commit the Concorde fallacy. Coaches play their high-salaried players more than is justified by their ability. Investors put significantly more money into failing ventures than into successful ones. Racetrack gamblers who have lost in early races are more likely to bet heavily on long shots in later races."
"Maybe they all just want to save face -- or maybe winning is less important than not losing." I was awake now, my mind functioning. "Or it could be that the same amount of money is perceived differently in different..."
"Yeah, yeah, it could be a lot of things, but the point is, people do this stuff all the time. We can't help it. It's like the Dartmouth students trying to find the pattern in the random lights. And it leads us to make the wrong decisions. But Dawkins claims that animals don't commit the Concorde fallacy."
"So let me guess. You're going to get a bunch of lab rats..."
"Pigeons, actually. And rig up a board with buttons to peck and lights that are wired to reflect the current prices of selected stocks."
"You're going to use pigeons to predict the stock market?"
"Specifically, to advise people when to drop an investment. I'm calling it Pigeon-Drop Investments."
I cradled my head in my hands. "Have another piece of toast, Corbett."
-- Michael Swaine