In addition to The Black Swan, trader and author Nassim Nicholas Taleb also wrote Fooled by Randomness, published in 2004. I’ve written before about the Problem of Induction, and how this relates to prediction in enrollment management. Here’s another tack on the topic.
Taleb has a respect for hedge fund manager and author (and champion squash player, it turns out) Victor Niederhoffer, who was educated at Harvard and the University of Chicago and whose Niederhoffer Investments was once the No. 1 hedge fund manager in the world, according to Marriott International, Incorporated (MAR). In 1997 he published the New York Times bestseller The Education of a Speculator.
This was the same year everything tanked. A bet on Thai bank stocks, and an incredible drop in the Dow Jones, led to what traders call — with all due poetry — a “blow up.” His investment firm was shuttered, but he did manage to climb from the rubble. Later, in 2007, he crashed again.
And still he goes on.
Despite the apparent failures, why does Taleb admire him so? Because Niederhoffer “derived his knowledge of the world from past data stripped of preconceptions, commentaries, and stories.” As I wrote about in an earlier blog on the “Black Swan” problem, we tend to look backwards and compose a narrative — a story — that helps us make sense of our failures. That helps us feel like things aren’t as random as they actually are.
In terms of higher education and enrollment management, this brings us again to the idea of “historical data.” Sure, we can look backward at various factors that might influence enrollment — test scores, GPA, zip codes and so on — and infer, intuitively, that certain of these prediction points just “make sense,” and so we should use them again. But how “testable” are those data points?
Taleb says that part of Niederhoffer’s dogma is “that any ‘testable’ statement should be tested, as our minds make plenty of empirical mistakes when relying on vague impressions … How many effects we take for granted might not be there?”
In Part 2 of this blog, I will look at a testable statement — one that, according to Taleb, “can be broken down into quantitative components and subjected to examination.”
By Sean Hill, Senior Content Writer, Capture Higher Ed