Mental Models: Make Better Decisions in Uncertainty
Expected value helps us make higher quality decisions. And feel better about them too.
Romance and regret taught me better decision making.
I had just finished a wonderful date. It’s the end of the night. I’m dropping my date off at her home.
She’s looking at me and smiling. I lean in and hug goodbye. As we separate slowly, we lock eyes. She’s still smiling at me. My mind is racing. My heart is thumping. I’m furiously debating what I should do next.
I give her a wave, say goodbye, and leave.
It was the most anticlimactic date I had ever been on. It was nothing like the movies. Unless the movie is an awkward comedy. I immediately regretted my decision.
Decisions with uncertain outcomes are hard. Uncertainty causes us to waffle back and forth. We doubt our thinking. Worse, we feel bad if we don’t get the outcome we were hoping for. Eventually, we shy away from decisions or default to lazy thinking.
Our lives are the sum of our decisions. Given the impact, it makes sense to build our decision making abilities.
There’s a way we can frame our decisions to not only get better outcomes, but to also help us to feel better about our decisions - regardless of the outcome.
Expected value thinking helps us think through decisions and uncertainty.
Expected Value Thinking
You’re about to head out for dinner. It’s damp outside. The weather report says 20% chance of rain. Do you bring an umbrella?
We weigh the possibility of rain and the outcomes of carrying an umbrella or not. When we run through our decision, we’re intuitively using expected value to make our decision.
The core of expected value is to look at decision options and outcomes. Each outcome is evaluated on two parameters: 1) The magnitude of the outcome, whether it’s positive or negative. 2) The likelihood of the outcome occurring given the decision made. Expected value comes from probability theory and provides a mental model we can follow in our lives.
For our umbrella example, our decision is whether we carry an umbrella to dinner or not. On one hand, if we don’t carry an umbrella and it rains, we’ll get wet and perhaps ruin clothes we’ve worn for dinner. On the other hand, if we carry an umbrella and it doesn’t rain, we have the annoyance of carrying an umbrella for the evening.
With expected value thinking, we focus on the outcomes with the biggest impact to us. How unhappy would I be if it rains and I get wet? How annoyed would I be to carry an umbrella all evening if it doesn’t rain? The magnitude indicates the importance of the outcome to us.
With outcomes and magnitudes, we look at likelihood of the outcomes. The weather report indicates a 20% chance of rain. We compare our two decision choices against the outcome impact and the likelihood of the outcome.
If I would get really annoyed with carrying the umbrella, it would make sense to not carry the umbrella, considering the annoyance and higher probability of no rain. If on the other hand, I would get extremely upset if I got wet and only mildly annoyed at carrying the umbrella, then I could be better to carry the the umbrella.
The key is to decide based on the outcomes and likelihood of each decision.
The choice to bring an umbrella has straightforward and low impact outcomes. Let’s apply expected value thinking to a higher stakes decision.
Let’s say you’d like to start your own business. You have a solid job already. The decision faced is whether you quit your job and pursue your own business, or stay in your current job.
Some of the outcomes are illustrated below.
The outcomes with the highest impact is a personal decision based our own situation and values. In this case, we identify failure or success of our business as significant outcomes. So is the regret of not starting a business. Now we compare the decisions with outcomes and probabilities together.
Notice we have high confidence outcomes for both decisions. If we don’t start a business we’re 100% sure we’ll regret the decision. On the other hand, if we start a business, we have a 80% chance of failure.
High confidence outcomes enable us to evaluate outcomes as certain. If the pain or cost of failure is extremely high, we may be best to not start a business. However, if the pain of regret (which in this case is certain) is higher than the pain of failure, we’d be best to start the business, despite the high risk of failure.
The best decision will be unique to each of us. The magnitude and likelihoods of outcomes are personal and subjective. The level of regret or pain from failure will differ among people of different values and experiences.
Insights from expected value thinking
Expected value thinking offers insights to improve our decision making skills.
Probabilities matter
We misjudge decisions if we gloss over probabilities of outcomes. Probabilities distinguish the possible and probable. Possible outcomes with low probability reduce the significance of the outcome. If there’s a 1% chance of rain, the worry of getting my clothes wet is low. Carrying an umbrella to “better be safe than sorry” is wasteful.
Probable outcomes warrant attention since we can expect the outcome to occur. There’s a high probability of failure if one chooses to start a business. In this case the failure outcome should be reviewed carefully.
We can tip the scales in our favor
We can influence the outcomes of our decisions in two ways: 1) Modify probabilities of outcomes in our favor. While starting a business has high probability of failure, we can reduce the probability of failure through activities such as working with mentors or studying other businesses. 2) Modify magnitude of the outcomes. For negative outcomes, we can examine for ways mitigate losses or “soften the blow.” For a business, we might have a backup direction in the event of failure.
Our data is often wrong
We often use wrong information to make our decisions. Estimates of impact of outcomes can be wrong. Humans exhibit an impact bias, where we overestimate the emotional intensity of future events1. We think the failure of a business will be more painful than it actually is. Outcomes may be non-obvious. Regret is an outcome we may not consider until we actually experience it after a decision has been made.
Expected value thinking provides a base to evaluate details relevant to the decision.
Why does expected value thinking matter?
We focus on the process, not the outcome.
When we shift to thinking with probabilities and outcomes, we recognize there is uncertainty in the result. We don’t have complete control over the outcome.
This empowers us to focus on the process of our decisions. Our process determines the quality of our decisions. Not the outcome. We have more confidence in our choices, and less worry when things don’t go as we hoped.
Higher quality, lower effort
Unstructured thinking can particularly effortful. We stir in circles of confusion and anxiety. Humans like the path of least resistance. We take shortcuts that lead us to worse decisions. Expected value thinking provides a model to guide our thinking.
Good decisions compound
Decisions and outcomes exist in a cycle. Better decisions lead to better outcomes, which lead to better decision opportunities. A manager who makes decisions with better outcomes 80% of the time instead of 70% will experience an outsized impact over time. 10% better judgement on decisions on the scales of thousands, millions, or billions of dollars matters. And it adds up quick.
“If you can be more right and more rational, you’re going to get nonlinerar returns in your life” - Naval Ravikant
..And therefore what?
Humans don’t like uncertainty. We also like to be lazy. Decisions with uncertain outcomes are a double whammy - effortful thinking with uncertain parameters. Anxiety creeps in and we make lazy decisions or shy away from decisions altogether.
Those who can make good decisions in uncertain conditions benefit greatly. A mental model for decision making builds our skills, enables us to lead, and to reach the outcomes we desire.
“If I make, like, three good decisions a day, that’s enough, and they should just be as high quality as I can make them” - Jeff Bezos
Years after my anticlimactic date, I’m on another date.
The date has gone well. It’s the end of the evening. I’m dropping my date off at her home. We exchange pleasant goodbyes. We’ve seen each other several times now. I want to kiss her goodbye.
My mind is racing. What do I do? How do I kiss her with suave? Does she like me? What if she rejects me? Will she think I’m a creep?
My date sees my blank stare. She looks confused. She turns to open the car door. In the split second before she leaves, I become keenly aware of the massive regret I’ll feel if I don’t act.
“Umm..” I mutter. She turns and looks at me curiously. I have no plan and no smooth lines. I sheepishly sputter, “Can I kiss you?”
She smiles, leans over, and gives me a kiss goodbye.
I was ecstatic. I experienced the joy of a date gone well, a kiss, and avoided massive regret. And I got to look forward to a future date.
Wilson, Timothy D., and Daniel T. Gilbert. "Affective forecasting." (2003).