Longtime readers will know that I stress the importance of randomness in investing and trading. Any investing or trading strategy will produce a number of losers. Sometimes these losing trades or investments will cluster together purely by chance; it can be difficult to distinguish these chance losing streaks from a declining in a trading system’s performance.
Those without statistics training may find it difficult to think in probabilistic terms and and often do not realize how random the world truly is. Following is a real life example of randomness in action. Take two bright, motivated, and somewhat unusual young men, just five years apart in school, both living in the Chicago metropolitan area. One of them, named Michael, attends a pretty good suburban high school. The other, named Franklin, attends the state math and science magnet school. When it comes time, both apply to a number of colleges, including the most exclusive and unusual college in the country, Deep Springs.
The application process to get into Deep Springs is the most demanding application to any undergraduate institution in the country. After writing over 20 pages of essays each, submitting recommendations and standardized tests, both Michael and Franklin are given the chance to visit Deep Springs where they face a second screening, intended to winnow the applicants further (this is early in 1999 for Michael and 1995 for Franklin). One is rejected, the other accepted. Here their paths diverge.
Franklin attends Deep Springs for two years and then transfers to Harvard. He graduates with high honors, marries, and then works for three years in a psychology research lab at Brandeis University. He then applies and is accepted to enter the Psychology PhD program at Washington University in St. Louis, under renowned researched Henry L. Roediger III.
Five years after Franklin is accepted by Deep Springs, Michael is rejected by Deep Springs and instead he begins college at Grinnell College in Iowa, a well-respected liberal arts college. Three months later he drops out and moves back home, working a retail job for 45 hours a week while taking classes full-time at the local community college. After three semesters he transfers into to a virtually unknown Quaker liberal arts school in Indiana named Earlham College. He studies for a year, takes a year off to work in France, then after another year he graduates with high honors. He applies immediately to graduate school, and he is accepted to the Psychology PhD program at Washington University in St. Louis, studying with Henry L. Roediger III. Not long after starting graduate school he marries his college sweetheart.
This is a true story of how two similar people end up in the same place despite wildly divergent paths (and their paths have since diverged again). Randomness is a powerful force and given the large number of uncorrelated events we all experience in our lives, coincidences are bound to happen and sometimes they are so prevalent that we become convinced that fate and destiny exist.
Never underestimate the power of randomness. When it comes to trading it is easy to see illusory correlation and find spurious correlations that do not continue and were simply the result of random chance.
5 thoughts on “Deep Springs: A lesson in randomness”
Congrats! Very well written post.
Could you please elaborate on where you see randomness in that real life example? Do you mean two boys, from Chicago era, landed in the same PhD program? According to me, there are only decisions made by humans. Sure, coincidences exist ex-post but not ex-ante.
That is a good question! There is a reason this post had languished in my drafts folder for years. I did not do a good job of explaining it (and I will add to the post if I figure out a way to do it), but the large coincidence of meeting someone with such a weird similarity in history makes it seem unlikely post hoc. But if we look at the set of all possible unlikely coincidences, the likelihood of any one unlikely coincidence is very, very high. Linking it back to trading, I would say it is another way of describing the data mining problem or the prevalence of spurious correlation.
Great post, Michael! It reminds me of the “birthday problem.” I can’t remember the exact numbers, but I think it was if you have 30 people in a room, the probability of two of them having the same birthday was something like 70 or 80%. People dramatically underestimate the probability of coincidence.
“Fooled by Randomness” is a great book on this topic. “How We Know What Isn’t So” is also a good one.
Yeah I really can’t recommend enough “Fooled by Randomness.” I need to read that other book.