Scenario Planning and Expected Value
Map possible futures and weight them by probability before deciding
Scenario planning is the practice of imagining multiple possible futures that could result from a decision, estimating the probability of each, and using those estimates to make more informed choices. Rather than predicting a single outcome and planning for it, you map the full tree of possibilities and prepare for the range.
The framework goes beyond simple brainstorming by introducing expected value -- the probability-weighted value of each scenario. A $100,000 grant with a 25% chance of success is worth $25,000 in expected value, while a $50,000 grant with a 70% chance is worth $35,000. Without probability weighting, you would prioritize the larger amount; with it, you would correctly prioritize the smaller but more likely one.
This approach also provides structural protection against resulting and hindsight bias. When you have mapped multiple futures in advance and estimated their probabilities, you are less likely to treat whichever future actually occurs as if it was inevitable. You can see it in context alongside the other futures that could have occurred, maintaining a more rational perspective on both your successes and failures.
- Every decision leads to multiple possible futures; planning for only one is planning to be surprised.
- Expected value (probability times outcome value) is more useful than raw outcome value for decision-making.
- The reason we do reconnaissance is precisely because we are uncertain -- imprecise estimates beat no estimates.
- Even a wide probability range like 20-80% is better than the implicit default of 0% or 100%.
- Mapping futures in advance protects against hindsight bias and resulting.
- Identify the decision and list possible futuresFor any significant decision, brainstorm all possible outcomes, both positive and negative. Do not just consider best case and worst case -- include the full range of intermediate outcomes. For each decision alternative, map the tree of what could happen next.Pro tipInvolve diverse perspectives in the brainstorming. Different viewpoints reveal scenarios that any individual would miss.WarningDo not stop at the most obvious outcomes. The scenarios you fail to consider are the ones most likely to catch you off guard.
- Estimate probabilities for each scenarioAssign a rough probability to each possible future. These will not be precise, and that is fine. The goal is to move off the implicit extremes of 0% and 100%. Even rough estimates dramatically improve planning. If two members of your group disagree strongly on a probability, have them argue each other's position.Pro tipStart with whether the scenario is more or less likely than 50%, then refine from there.WarningResist the urge to avoid guessing because you cannot be precise. You are already implicitly guessing when you choose one action over another.
- Calculate expected valuesMultiply each scenario's value (positive or negative) by its probability to get the expected value. Sum these across all scenarios for each decision alternative. The alternative with the highest total expected value is often the best bet, though other factors like risk tolerance may apply.Pro tipA $200,000 opportunity with a 10% chance of success ($20,000 expected value) is less valuable than a $50,000 opportunity with a 70% chance ($35,000 expected value).
- Plan responses for key scenariosFor the most likely and most impactful scenarios, develop contingency plans. What will you do if the best case happens? The worst case? The most likely case? Having pre-planned responses makes you nimble rather than reactive when the future unfolds.Pro tipConnect scenario planning to Ulysses Contracts by identifying scenarios where you are likely to make emotional decisions and precommitting to rational responses.
- Close the feedback loopAfter the outcome is known, compare what actually happened to your scenario map. Were your probability estimates reasonably calibrated? Did you miss important scenarios? Use this to improve future estimates. This is how prediction accuracy improves over time.Pro tipASAS improved their grant estimates dramatically by going back to both successful and unsuccessful grantors to understand why decisions were made.WarningDo not fall into resulting when closing the loop. A low-probability scenario occurring does not mean your estimate was wrong.
After-School All-Stars was prioritizing grant applications by face value rather than expected value. A $100,000 grant with 25% probability was prioritized over a $50,000 grant with 70% probability, despite the smaller grant being worth $10,000 more in expected value ($35,000 vs. $25,000).
Carroll's pass call gave Seattle three chances to score instead of two. The probability of an interception was only 2-3%. An incomplete pass (most likely non-touchdown outcome) would stop the clock and still leave two running plays. A run that was stopped would force use of the last timeout, reducing to just one more play.
Duke adapted this from poker, where expert players routinely consider every possible opponent response (fold, call, raise) and their respective probabilities before making a bet. She then applied it in her consulting work with After-School All-Stars (ASAS), a nonprofit that was struggling with budget planning because they treated every grant application as worth its face value rather than its probability-adjusted expected value. The transformation in their planning was immediate and substantial.