THE IDEA
Two very different kinds of not-knowing
When you roll a dice, you don’t know what number will come up. But you know it’ll be between one and six, each equally likely. That’s risk. It’s quantifiable. You can calculate the odds, price it, insure against it, and make rational bets.
When a new technology appears and you’re trying to figure out what it will do to your industry, you’re dealing with something different. You don’t know the outcomes. You don’t know the probabilities. You may not even know the full range of what could happen. That’s uncertainty. It’s not quantifiable. You can’t calculate the odds because you don’t have the inputs for the calculation.
The distinction was formalised by economist Frank Knight in 1921, and it remains one of the most underappreciated ideas in decision-making. Most of the tools we use - risk registers, probability matrices, expected value calculations, insurance - work for risk. They assume you can list the possible outcomes and assign likelihoods. Under genuine uncertainty, those tools give you a false sense of security. The spreadsheet looks rigorous. The confidence is unearned. The most important decisions - career changes, market entries, relationship commitments, technological bets - sit squarely in the uncertain category, not the risky one.
IN PRACTICE
When the odds don’t exist
An insurance company prices car insurance with remarkable precision. They know the accident rates, the claim distributions, the demographic patterns. The risks are quantifiable because there’s enough historical data to calculate them. This is risk management, and it works because the future resembles the past closely enough for the statistics to hold.
A venture capitalist evaluates a startup with a genuinely novel product in a market that doesn’t exist yet. There’s no historical data. The comparable companies aren’t comparable. The projections are stories, not calculations. The VC who treats this like a risk problem - running financial models to three decimal places - is doing theatre. The honest assessment is: this might change everything or it might be worthless, and nobody knows the probability of either. The smart bet isn’t based on odds. It’s based on the size of the opportunity, the quality of the team, and the ability to absorb the loss.
A family decides whether to move to a new country. They can research the cost of living (risk - quantifiable). They can’t research whether they’ll be happy there (uncertainty - not quantifiable). They can’t know how the move will affect their children’s development, their marriage, their sense of belonging. These aren’t risks with calculable odds. They’re genuinely unknowable. The decision has to be made on different grounds - values, instincts, willingness to adapt - because the analytical tools don’t have inputs.
WORKING WITH THIS
Choosing the right compass
The first step is honesty about which kind of not-knowing you’re facing. If you can list the possible outcomes and roughly estimate their likelihood, you’re dealing with risk. Use the tools designed for risk: probabilities, expected values, diversification, insurance.
If you can’t - if the outcomes are genuinely unknown, or the probabilities are guesswork disguised as analysis - you’re dealing with uncertainty. Different tools: small experiments that reveal information. Optionality - keeping doors open rather than committing to a single path. Resilience - the ability to absorb surprises. Heuristics and principles rather than detailed plans.
The most dangerous error is treating uncertainty as risk. It feels more professional to put numbers on things. A risk matrix with colour-coded squares looks more serious than saying “we don’t know.” But unearned precision is worse than acknowledged ignorance, because it closes down the very adaptive responses - flexibility, experimentation, humility - that uncertainty demands. When you truly don’t know, the honest response isn’t a better spreadsheet. It’s a different way of moving forward.
THE INSIGHT
The line to remember
Pretending uncertainty is risk doesn’t make it manageable. It makes you confident about things you have no right to be confident about.
RECOGNITION
When this is in play
You’re dealing with uncertainty rather than risk when the possible outcomes can’t be fully listed in advance. When someone’s financial model has precise probabilities for unprecedented events. When a plan assumes the future will behave like the past in a situation where the rules are changing. When the phrase “worst-case scenario” appears in a document and you suspect the actual worst case hasn’t been imagined yet. When the decision that matters most is the one where the spreadsheet is least useful.