Being notionally correct beats false precision (Part 2)

In yesterday’s post, I argued that false precision in decision making, in the form of unreliable data or faulty conclusions, can create serious problems in all arenas of life.

So what’s the alternative? The alternative is an approach in which, in the absence of truly reliable data, decisions are based on solid notions. This approach is characterized by the following:

• First principles: Relying on first principles to derive what is a solid first assessment of the situation, without relying on any external data.
• Healthy skepticism: Approaching data with a healthy skepticism and approaching authority with a healthy iconoclasm.
• Combining the inside view with the outside view: This is a Daniel Kahnemann concept — it means that after making one’s own assessment, the decision maker selectively pulls in an outside view that is based on a class of roughly similar previous cases.

Let me give you three concrete examples where this approach has changed how I look at the world:

• Health and medicine: My approach to health is a barbell approach — On the one hand, if you are facing an acute health crisis, you often have no choice but to rely on the healthcare system. And modern medicine is actually very good at dealing with these situations (e.g. surgery, antibiotics or rapid responses to heart attacks). On the other hand, when it comes to avoid getting sick in the first place, my approach is one of first principles: Avoid things that have not been around for at least hundreds of years unless there is really solid data. This includes most pharmaceuticals, foods, supplements or even cosmetics. The reason is that in the absence of really solid data, you are actually exposing yourself to greater risks with often questionable benefits. This is the definition of a bad bet.
• Company metrics: In my past career at Google I was involved in defining metrics to measure performance for a large subset of the business organization. In many ways, while the intention was a good one, this was actually a fantastic example where false precision causes more problems than benefits. The reason is that it was very difficult to find metrics that were solid proxies for actual performance. But once established, these metrics were taken very seriously causing a lot of unnecessary internal arguments, distress and probably a less successful attribution of performance. Here is a good take on why measuring performance should be first and foremost a manager’s job and should include much more than the false precision of a few metrics.
• Portfolio construction and asset management: I’m currently attending a conference for asset managers in which one issue is becoming very clear: there is a tremendous over-reliance on time series data to assess the future risk and return of an investment or asset class. This approach works well until it doesn’t. As the saying goes “Past performance is no guarantee for future results”. This is particularly true when an economy is undergoing a true phase shift, as I believe is currently the case. This environment requires a much greater reliance on first principles thinking and a true openness to a change in patterns, while not losing sight of the data.

Good luck making your decisions out there and please let me know your feedback!

Thomas Moran’s “Green River Cliffs, Wyoming” (1881)