Expected Value Mindset
Weight every possible outcome by its probability before placing a bet
- Difficulty
- Moderate
- Time to result
- ~weeks to results
- Steps
- 6
- Confidence
- 98%
An expected value mindset evaluates a choice across its full range of possible outcomes. Identify the major scenarios, estimate their probabilities, attach a payoff or cost to each, and combine them into a probability-weighted value. The objective is not to predict one certain future but to locate situations where repeated exposure would produce more gains than losses. Decision quality must therefore be separated from the single outcome that happens next. A favorable bet may lose, while a poor bet may win through luck. Practical use also requires attention to estimation error, downside severity, and bet size; a mathematically positive opportunity is unsuitable if one loss creates ruin or if its probabilities are unsupported.
Origin
Extracted from Masters in Business during Michael Mauboussin’s discussion of probabilistic fields, sports drafting, contracts, and investment odds.
Core principles
- 01Judge decisions across repeated possible worlds, not one realized outcome
- 02Probability and payoff must be assessed together
- 03A good decision can still produce a bad result
- 04Edge comes from favorable odds rather than certainty
- 05Downside magnitude matters as much as win frequency
How to run it
- 1
Map outcomes
List the materially different ways the decision could resolve, including adverse and extreme cases.
Pro tip Use ranges when the outcome space is continuous.
Watch out Do not omit uncomfortable outcomes simply because they appear unlikely.
- 2
Estimate probabilities
Assign probabilities or probability ranges using base rates and relevant evidence.
Pro tip Check that mutually exclusive probabilities sum to one.
Watch out False precision can conceal weak assumptions.
- 3
Value payoffs
Estimate the gain, loss, or utility associated with each possible outcome.
Pro tip Include opportunity costs and second-order consequences.
Watch out Do not compare outcomes using inconsistent units.
- 4
Weight and compare
Multiply each payoff by its probability and sum the results for each available choice.
Pro tip Compare the result with doing nothing and with the next-best alternative.
Watch out Positive expected value does not eliminate outcome variance.
- 5
Control exposure
Size the commitment so that adverse outcomes remain survivable and repeatable.
Pro tip Reduce size when probability estimates are fragile.
Watch out Never accept a risk of ruin merely because the arithmetic mean is positive.
- 6
Review the process
After the outcome, evaluate the assumptions and reasoning separately from whether the decision won or lost.
Pro tip Maintain a decision journal containing the original probabilities.
Watch out Do not infer decision quality from one result.
In the wild
A general manager groups candidates by age, measured ability, and comparable career histories. The team estimates each player’s probability of becoming a starter, reserve, or failure and values those outcomes against salary and draft cost.
→ The contract and draft decision reflect the full distribution rather than a scout’s single narrative.
An investor estimates a forty percent chance of gaining £300 and a sixty percent chance of losing £100. The probability-weighted gain exceeds the expected loss, but the position is kept small enough that repeated losses remain survivable.
→ The investor takes a favorable but controlled bet without pretending the next outcome is certain.
Common mistakes
Predicting only one future
A point forecast hides the range of outcomes and prevents explicit assessment of uncertainty.
Ignoring payoff asymmetry
High win probability can still be unattractive when rare losses are catastrophic.
Scoring by the result
Treating every win as proof of skill and every loss as proof of error rewards luck rather than process.
Is it for you?
Best for
Decision-makers choosing among uncertain investments, contracts, hires, strategies, or other repeatable bets.
Not ideal for
Choices involving unacceptable ruin, non-negotiable ethical constraints, or consequences that cannot sensibly be valued.
From the transcript
“what you're really trying to do is figure out situations where you have some sort of advantage or some sort of an edge, where if…”
“I know the range of outcomes, I know the probabilities or some sense of those things, and I'm only gonna bet, by the way, when…”
From the episode
Interview With Michael Mauboussin: Masters in Business (Audio)
Michael Mauboussin