Robert Rubin on Bayesian Thinking

^z 9th May 2023 at 7:09am

"I’m a Former U.S. Treasury Secretary. Here’s How I Make Hard Decisions." is an anecdotal NYT op-ed essay summarizing key points in Robert Rubin's soon-to-be released book, The Yellow Pad: Making Better Decisions in an Uncertain World. The journey is good, though the route is unsurprising. In a nutshell, Rubin counsels:

  • list potential outcomes
  • estimate probabilities
  • decide based on costs and benefits, weighted by likelihoods

All quite obvious, and nontrivial to do in real life. Echoing Benjamin Franklin's decision algorithm, Rubin writes:

At the heart of my own approach is “probabilistic thinking,” the idea that nothing is 100 percent certain and that everything is therefore a matter of probabilities. Whether my choices would affect a few people or millions of people, my preferred tool for applying probabilistic thinking has always been the same: a simple yellow legal pad. On my yellow pad (or more recently, my iPad), I’ll list possible outcomes in one column, and then my best estimates of the probabilities associated with those outcomes in another.

My goal has never been to quantify every aspect of every decision; that would be impossible. Instead, my yellow pad has become both metaphor and means, a way of applying a questioning mind-set and incorporating probabilistic thinking into the real world. There are, of course, decisions throughout my life that looking back I should have made differently. But the yellow pad has served me well, allowing me to think in disciplined ways about risks, probabilities, costs and benefits, and substantially increasing my odds of making the best possible choice. What’s more, I believe the yellow-pad approach can be beneficial for everyone.

For example, applying probabilistic thinking to real-world events changes the way one thinks about risk. Too often, decision makers trying to anticipate a risk focus on a single potential outcome, or perhaps a small handful of outcomes. Probabilistic thinkers, on the other hand, recognize that risk is a wide range of possibilities.

Rubin goes on to emphasize the importance of trade-offs among multiple competing goals, of embracing nuance and complexity, and of evaluating past decisions based not just on their results but "... also analyzing the judgments that led to the decision and took place before those outcomes occurred." All excellent points!

Bottom line: try to be a Bayesian thinker, basing your judgments on evidence and updating them when new data come in. Remember:

Beliefs are Knobs,
not Switches!

(cf Mantra - Beliefs Are Knobs, Not Switches (2017-07-03), ...) - ^z - 2023-05-09