The Danger of Easy Skepticism

I remember taking a class in business school, where we had to evaluate the financial prospects of startups. The startups were real, but the names were changed and in industries obscure enough that we rarely could guess the true identity without cheating.

Given limited information and numbers, we were each given one company and asked to present our analyses. Soon, however, a dominant strategy emerged: nearly every presentation predicted failure.

For the incentives of the classroom, that made sense. Most new startups fail, and for many different reasons. It’s easy to find a reasons for failure, and when you predict failure, you’re usually correct. As a result, the class taught us to practice skepticism, find reasons why things might go wrong.

But real life isn’t like the classroom. An investor with ten companies may see several fail, a few break even and only one earn enough to make up for all the past losses. Skepticism, in this instance, is easy but also not terribly useful.

Training Optimism

This doesn’t just apply to startups. More formally, it applies to any situation where there is an asymmetric prior probability of success and equal (or greater) asymmetry in the results of that success.

Put simply: When failure is cheap, but frequent, and success is lucrative, but rare, it pays a lot more to recognize successes.

You can imagine the situation in a 2×2 matrix:

1. Failure cheap, failure rare: No-brainer optimism.
2. Failure pricy, failure rare: Trained skepticism.
3. Failure cheap, success rare: Trained optimism.
4. Failure pricy, success rare: No-brainer pessimism.

When success is likely and profitable, we all feel optimistic. Similarly when failure is likely and costly, we’re all pessimistic. It’s the latter two situations, where rare events also have uneven payoffs, that our intuitions often betray us.

For a good talk about many situations where cultivating trained skepticism is a good idea, Nassim Taleb’s The Black Swan, is a great book. Other websites like OvercomingBias and LessWrong, provide a lot of tools and examples of trained skepticism.

I feel trained optimism is a somewhat less discussed skill, so I want to highlight it here.

Is Trained Optimism Underrated?

Giving an overly optimistic prediction or idea, and having it be wrong, makes you look gullible and foolish. Giving an overly pessimistic prediction or idea, and having it be wrong, is generally forgiven as being cautious. Therefore, if your goal is only looking intelligent, and you have no other stake in it, skepticism is the more suitable strategy. Indeed, that’s what I observed from our classroom experiment: almost nobody was willing to bet their flawed startup would succeed.

But what if you have real stakes and not just your conversational reputation? Not every domain of life warrants trained optimism, but some certainly do.

I think a lot of personal and professional projects largely fit under this category. There’s some minimal opportunity costs and extra effort, and most the time they won’t be great successes. But some of the time they will, so the person who keeps trying and experimenting will beat the skeptic in the long run.

Learning new things is almost always this case. I’ve learned many things that have never served a practical purpose in my life, but often the few that do are quite unexpected. Doing the MIT Challenge, for example, I found the classes on probability and logic more useful than many of the programming classes, something I hadn’t expected.

Optimism as a Skill and a Mood

Bear in mind, this trained optimism is about trying to more carefully pick winners (rather than losers) not a blanket policy of good feeling about every possible action one could take. Also, if the size of success is still dwarfed by its improbability, it’s probably still a bad decision.

In this sense, optimism is a skill. It’s focusing on improving your ability to pick good project candidates, business ideas, habits to follow and not worry too much about the ones that fail. It also means accepting slightly less accurate beliefs, for the extra payoff big wins can create.

But optimism is also a mood, not just a skill. I don’t think one can mechanically implement trained optimism (or trained skepticism, in the situations where that is better suited) without changing your mood about a certain area of life. When hard numbers aren’t available, we rely on moods as heuristics for processing the world and biasing our thinking in a favorable direction.

Being able to pick winners requires looking for ways things can go right and not just the ways things can go wrong. It requires tuning yourself to see the subtler positive signs instead of the overwhelming negative ones. In that sense, optimism is a cultivated emotion, not just a calculation, and at least some of the time, the smarter bet.