Can you assign probabilities to the future?

edit: this spurred further thought into my understanding of uncertainty, which turned into Definitions: Information, Predictions, Risk, Randomness, Markets..

image

When I come across strongly conflicting views from thinkers that I otherwise respect, I pay attention. Here are two on risk and assigning probabilities to various possible outcomes:

From What I Learned Losing a Million Dollars (Jim Paul):

Risk is defined as the possibility of suffering a loss. In a certain world, we wouldn’t have the potential for loss that is risk. It is the uncertainty of the future that triggers the negative psychological processes described. But it is precisely because the future is unfathomable that there are always buyers and sellers in every market. Most people erroneously assign a numerical value to chance, which confuses risk with probability. In the markets we are talking about unique, non repeatable events, so we can’t assign a frequency probability to their occurrence. All you can really determine is the amount of your exposure and losses, not predict profits.

From The Dhando Investor (Mohnish Pabrai):

Wall Street sometimes gets confused between risk and uncertainty, and you can profit handsomely from that confusion. Wall Street just hates uncertainty of any kind, and punishes prices accordingly.

Jim Paul was a commodities futures trader who blew up in the 1980s. The book is a fantastic treatise on the layers of misconceptions and self-deception that prevent people from letting go of their losing trades. Several key ideas of Nassim Taleb’s Antifragile are sourced from here, including “some people want to make money, others want to be right” and the lumber trader who took 10 years to realize that green lumber meant it was freshly cut, not actually painted green.

Mohnish Pabrai is a value investor and long-time fan of Warren Buffett. Part of Pabrai’s investment process is finding distressed companies, coming up with the available scenarios and assigning each a probability and general upside or downside range. He uses that, coupled with the Kelly criteria, to determine if an opportunity is a sure bet, and has profited handsomely from this strategy.

These are two fundamentally different views of the world - Pabrai depends on probability-scoring events, which Jim Paul considers impossible. To some degree, I think Nassim Taleb would as well - true risk assessment is impossible because the probability of extremely unlikely events is impossible to compute. Often, probability assigning is also difficult because there is no way to incorporate a feedback loop into the projections. All one can depend on are historical examples, which don’t predict the future.

I spent several days wrestling with this contradiction. My conclusion: prophesy concerns itself with the unknowable, while prediction does not.

The crucial piece here is that prediction is having some probability distribution to represent a view of the future. Thus one key output to creating such a view is the confidence interval.

Some things can actually “be predicted,” which is to say the confidence interval can allow for reasonable certainty (i.e. high likelihood) of a specific outcome. The vast majority of things, however, should have such a wide confidence interval that it is virtually useless to try and plan for them.

Wisdom is the ability to distinguish one’s confidence intervals (Buffett’s circle of competence). Wisdom is also understanding your exposure when you confuse the two, and taking out an insurance policy in case you are wrong (antifragility).

The great investors are not those who correctly assign probabilities to potential future outcomes. Remember, there is no way to ever verify if those probabilities were accurate or not. But, like Jim Paul says, some people want to be right, others want to make money. A successful investor is one who survives because she knows when it is a waste of time to assign probabilities in the first place, because some things are unknowable. She finds the very, very small set of outcomes that can be predicted, the future that is not Jim Paul’s “unfathomable.”

Unless you have divine inspiration, don’t prophesy. And definitely don’t bet your (or anyone else’s) money on your prophesies. Instead, spend a ridiculous amount of time trying to understand what is the realm of prophesy, and what isn’t. Assume that almost all of the future is probably unknowable. But when you find a small pocket of knowable future, to quote Pabrai, make few bets, big bets, infrequent bets.