In Ep. 491: Brian Christian with Michael Covel on Trend Following, Michael Covel discusses concepts from the book Algorithms To Live By: The Computer Science of Human Decisions with one of its co-authors, Brian Christian. As an introduction to the text, Christian says the ideas in the book represent an opportunity to incorporate computer science principles into to our own everyday decision making process. How would a computer scientist solve this particular problem? is essentially what it asks the reader to consider when confronting decisions in life.
Covel asks Christian to provide a simple explanation for a concept many non-computer science people tend to regard as something to fear, a complicated idea beyond their scope of knowledge, the Algorithm.
A formalized definition of algorithm from merriam-webster.com is:
a procedure for solving a mathematical problem (as of finding the greatest common divisor) in a finite number of steps that frequently involves repetition of an operation; broadly : a step-by-step procedure for solving a problem or accomplishing some end especially by a computer.
Christian and Covel add additional, more common and historical context explaining they can be thought of as a way to work out some of the steps and procedures commonly done with pencil and paper, but then formalized in computer science terms. Christian says:
We associate the algorithm with computer science, but its really a set of instructions to solve any problem. Its much broader than computers – we all learn algorithms in school math, and a recipe is even an algorithm. Math procedures have been found on clay tablets, the first and oldest examples in the human race.”
Covel goes on to provide an example of hiring an assistant where he posted an ad on Craigslist and received about twenty applicants. As he interviewed each of them, he wasn’t entirely sure if one particular candidate would be best suited for the job or if one scheduled to be interviewed later might be better. This common problem has come to be known as the Secretary problem. In this formulation, the subject is hiriing for a position and N candidates show up in random order for interviews at which point you must hire on spot or dismiss the candidate. In the Secretary problem formulation, you wouldn’t have the option call back ones you have dismissed. This is also known as the optimal stopping theory. How do you hire without knowing who the best candidate would be if you interviewed them all? What if you choose a good one and but maybe the best is still out there? This algorithm formulation is found throughout the course of everyday human decision making, for example when buying Real Estate.
They go on to discuss how computer scientists formulate the problem which is quite interesting, worth a listening to the whole episode for more details.
There is one additional principles discussed during the show that resonated, that of the Multi-armed bandit:
“problem in which a gambler at a row of slot machines (sometimes known as ‘one-armed bandits’) has to decide which machines to play, how many times to play each machine and in which order to play them. When played, each machine provides a random reward from a probability distribution specific to that machine. The objective of the gambler is to maximize the sum of rewards earned through a sequence of lever pulls”. The everyday problem that Covel and Christian discuss here is that, in order to properly exploit the probability of situations, we need to have a certain amount of data to make an informed decision.”
Ultimately, no amount of data gathering will ever provide 100% certainty, so where is the point at which we leap off and actually make the decision? In mathmatics and computer science, there is a optimal mathematical formula, and probably best to read the book for the details, but what one consideration the two present is a “regret minimization framework“. They offer the example of Jeff Bezos leaving a lucrative finance job to pursue the seemingly crazy idea to start an online book store. The regret minimalization idea here is that when you are faced with a decision and project forward to place in time where you can look back at the decision (something I have often called the “deathbed perspective”) would you regret not taking the chance? Covel sums it nicely with something to the effect of: “When you look back at your life, would you rather have worked in the same place for 50 years and collected the gold watch at retirement or take the risk and follow your passion?” The answer is clear stated in this context.
Make sure to check out the podcast and book.