What's Your Prior?

The Preacher from 1700’s England who Taught Wall Street how to Count

June 05, 2020 Damian Handzy Season 1 Episode 1
What's Your Prior?
The Preacher from 1700’s England who Taught Wall Street how to Count
Chapters
What's Your Prior?
The Preacher from 1700’s England who Taught Wall Street how to Count
Jun 05, 2020 Season 1 Episode 1
Damian Handzy

An introduction to the ideas behind this podcast, to the Reverend Thomas Bayes, why this podcast is named "What's Your Prior," and to your host, Damian Handzy.  

This is a good place for an intro, especially if you've joined in the middle of the series, to get some background on the podcast series, before going to the most recent episode.

Mentioned in this episode:

Show Notes Transcript

An introduction to the ideas behind this podcast, to the Reverend Thomas Bayes, why this podcast is named "What's Your Prior," and to your host, Damian Handzy.  

This is a good place for an intro, especially if you've joined in the middle of the series, to get some background on the podcast series, before going to the most recent episode.

Mentioned in this episode:

Damian Handzy:

Not long after Thomas Bayes , a Presbyterian minister died in 1761, a friend of his found an astonishing paper that the preacher wrote that would immortalize his name. Published in 1763, the paper explained how to incorporate new information into our beliefs and how to update our position on any topic. One of the mathematicians who worked on it after Bayes' death actually said that Bayesian theorem... "is to the theory of probability what the Pythagorean theorem is to geometry. But as time passed, the paper was largely forgotten, until the modern era, when the advent of computers made it really useful. Today Bayesians statistics are used in medicine, economics, technology, and a bunch of other fields. Bill Gates even credited Microsoft's competitive advantage to using Bayesian networks. Possibly most significantly, Alan Turing himself used Bayesian information to crack the enigma code and stop Nazi Germany in its tracks.

Introducer:

Hello and welcome to "What's Your Prior?" The podcast for the adaptable investor with your host, Damian Handzy.

Damian Handzy:

I'm Damian Handzy, your host on what I hope will be an interesting, sometimes surprising, and maybe even entertaining journey while we explore how seasoned professionals make investment decisions. Now, this is episode zero. It's the episode that gives me a chance to introduce some of the ideas behind this podcast and to set the stage for what you can expect, if you choose to listen. I'm going to get back to the story of how Bayes's work helped the allies win the war towards the end of this first episode. But first I want to kick things off explaining what this podcast series is all about. Let's start with who I am . I describe myself as a recovered physicist. I've worked in financial technology and in investment management for the past 25 years, I've been a consultant and entrepreneur and even a CEO. Currently, I'm the head of research and I'm the Chief Commercial Officer for Style Analytics, which provides investment style and factor analysis for professional investors all around the world. Our reports, our software, they're used by about half of the 100 largest fund managers in the world. And I get the privilege of interacting with them, learning from them and helping them make sense of financial markets. I grew up in a house where one parent was a mathematician turned physician and the other parent was a psychologist. While my wife says that sounds like a setup for an 80 sitcom, what it really did was prime me to value both numerical and analytical ways of thinking, but it also gave me a deep respect for human and emotional ways of making decisions. And my career in financial analytics has allowed me to, well, at least partially, to satisfy an absolutely incredible curiosity about making better investment decisions. So this podcast is for professional investors of all types. And while I have a quantitative background, this is not a podcast for quants. And while I'm fascinated by behavioral economics, it is not a podcast for psychologists. What it is, is a podcast that's going to try to get us to examine our assumptions and to look for multidimensional answers to questions of why markets do what they do. Each episode I hope is going to take a fresh perspective on some aspect of the markets. Now, my own point of view is that understanding financial markets isn't rocket science. It's actually a lot harder than that. I left nuclear physics and I started a career in financial markets 25 years ago partially because this field is actually not yet well understood. Financial markets cannot be solved with the math that we use to put probes on Mars or to measure the composition of atmospheres around planets light years away. Financial markets involve nonlinear feedback loops. They adapt to external events and they evolve. They're what scientists call complex adaptive systems. It's pretty much the opposite of what classical economics teaches us with its ideas like equilibrium. Now I flatly and vehemently reject the notion of equilibrium in markets. Financial markets, like economies, are dynamic. They're constantly adjusting to all their moving parts. Constantly. There is never equilibrium. I like this topic, not only because it's a challenge to understand. I mean, this stuff is hard, but it's also important to people around the world since virtually everyone in developed nations invests in stock markets, through their retirement funds, primarily this stuff matters to a lot of people. It matters to you, to me, to our parents, to our neighbors and people that town next over. All right, so how does Bayesian thinking fit into all this? Well, one of the most insightful questions I've ever heard someone asked, what would it take for you to change your mind? In other words, what evidence would you need to realize you've been wrong about something? That is exactly what Bayes's paper and his formula address. Bayes identified three key parts in the process of evaluating beliefs in the light of new information. Number (1), how confident are you in what you believe? That's your starting point. Number (2) - new evidence that you've learned either for- , or against-, your starting point. What Bayes's formula does is takes those two things and it tells you (3) how confident you should be now that you've taken the new evidence into account. The term that the Bayesian approach uses for your starting point, your belief before you get t he new information, the word that Basyesian people use is a "prior", as in, "PRIOR to getting the evidence, this is what I thought." Now, I think of a "prior" as the assumptions o r the biases that I bring with me - it's how I see the world. This podcast is named "What's Your Prior." It's kind of an homage to Bayesian thinking, and it's meant as a probe into what we believe about how financial markets work. It also asks us to examine our assumptions, our biases, our priors, in evaluating the convictions of our beliefs. Now each episode of this podcasts, the plan is it's going to focus on a particular aspect of financial markets and we'll likely have at least one guest to help explore that topic throughout. I hope to challenge some conventional thinking to invoke complexity science, where it's appropriate and to bring some Bayesian thinking to these topics, especially in identifying our own biases and our priors. All right , so how did Bayes influence Alan Turing in his effort to break the enigma code? Alan Turing developed what he called a crypto-analytic process, based on the Bayesean approach that used sequential estimated conditional probabilities of the most likely settings on the Enigma machine. This let Turing guess a sequence of letters in any Enigma message based on how frequently those letters appeared in other Enigma messages that they already cracked and on the likelihood of specific German words being used in wartime correspondence. This, in turn, allowed him to severely reduce the number of wheel settings that were needed to be tested to crack each day's code. Without this, it simply took too long to crack the code and the future encrypted messages using the new code. So using this Bayesian and approach, he was able to crack the enigma code efficiently. Now, another example of Bayesian thinking is applicable to all of us today because of COVID. Suppose there's a medical test for a rare disease that has a 99% accuracy rate. I mean, 99% of sick people test positive and 99% of healthy people test negative for the disease. And also suppose that the disease is kind of rare. It only affects 1% of people. Now, if I go get a test and the results come back positive, what are the chances that I actually have the disease? The intuitive answer is 99% but Bayes's formula says, no, the answer is 50%. That's right. There's only a 50 / 50 chance of having this disease. Even after testing positive from a test that has a 99% accuracy rate. Now there are plenty of explanations to understand this intuitively all over the web. So I'm not going to try to offer one here. If you really want one, just Google Bayesian medical example, and a bunch will pop up. Anyway, that's the power of the Bayesian approach. Now, I think the the economist John Maynard Keynes, he really summarized Bayesian thinking perfectly when he said, "when the facts change, I change my opinion. What do you do, sir?" And with that, I look forward to recording my first episode, which is going to tackle the question of what's been going on with global equity markets during COVID and before that, but specifically why Value investing has done so incredibly poorly in the past few years. I'm going to have two guests from the hedge fund powerhouse, Mount Lucas Management - their CEO, Tim Rudderow and their portfolio manager, David Asbell. I hope you'll join me. Until then, I asked you to join me in thinking about our assumptions, our priors, about how the market should behave due to the COVID pandemic.