How to estimate what may happen next without pretending your hunch came from Mount Certainty.

Forecasting is the discipline of making explicit, time-bound predictions under uncertainty. It helps you say not merely “this might happen,” but “I estimate a 35% chance this will happen by this date, and here is what would change my estimate.” That little sentence is already better than most public debates, which is both impressive and depressing.

This page teaches practical forecasting: base rates, reference classes, decomposition, probability estimates, updating, calibration, and Brier-style scoring. The point is not to become psychic. The point is to become less vague, less overconfident, and more willing to learn from reality when it sends back your prediction with red ink on it.

 

 

 

Quick navigation

 

Best used for

    • Planning under uncertainty.
    • Estimating risks and opportunities.
    • Reducing overconfidence.
    • Tracking whether your judgment improves.
    • Making time-bound predictions that can be reviewed.

 

5-minute version

Use this when the problem is pressing and you need the fastest responsible version of the method. Not perfect, but better than sprinting into a decision while waving a flaming assumption.

    1. Turn your concern into a clear question with a deadline.
    2. Find a base rate or reference class.
    3. Break the question into smaller drivers.
    4. Give a probability, not just a mood.
    5. Write what evidence would raise or lower your estimate.
    6. Review the forecast after the deadline.

 

 

30-minute careful version

Use this when the issue matters enough to deserve a slower look. Thirty minutes of structured thinking can prevent thirty months of cleanup, which is apparently a bargain humans keep trying to avoid.

    1. Define a forecast question with a clear outcome and timeframe.
    2. Find relevant base rates or comparable cases.
    3. Decompose the forecast into key drivers and constraints.
    4. Estimate probabilities for each driver where possible.
    5. Assign an initial probability and explain why.
    6. List evidence that would update the probability up or down.
    7. Set a review schedule.
    8. Score the outcome later using simple accuracy review or Brier-style scoring.
    9. Keep a forecast journal to improve calibration over time.

 

 

Vignette: The 90% confident disaster

A manager says there is a 90% chance a project will launch on time. Someone asks for the base rate. Similar projects launched on time only 45% of the time. The manager explains that this team is different, which is the mating call of planning fallacy.

The team decomposes the forecast: staffing, review delays, technical dependencies, stakeholder approval, and budget. The revised forecast becomes 55%, with three specific indicators that would raise confidence. The project still might succeed, but now the forecast has stopped wearing a fake mustache.

 

Practice: apply this to one of your three current problems

Write down your three most important current problems. Pick one. Then apply the prompts below. Do not merely admire the tool from a safe distance like a museum visitor staring at a fire extinguisher.

    1. Write one prediction you care about.
    2. Make it specific: what exactly will happen, by what date, according to what observable standard?
    3. Find a base rate or at least a comparable case.
    4. Assign a probability.
    5. Write two things that would increase your probability and two that would decrease it.
    6. Review your prediction later and write what you learned.

 

Common mistakes

    • Making forecasts without deadlines.
    • Using words like “likely” without numbers.
    • Ignoring base rates because your situation feels special.
    • Failing to update after new evidence.
    • Not recording forecasts, which lets memory quietly rewrite history like a corrupt historian.

 

 

AI Prompt Support Module

Use AI as a thinking partner, not as a priest, judge, or magical vending machine for certainty. First write your own answer. Then ask AI to challenge, improve, and stress-test it.

Make my forecast specific

Turn this vague prediction into a clear forecast question with a measurable outcome and deadline: [prediction]. Suggest possible base rates and what evidence I should gather before assigning a probability.

Decompose a forecast

Help me decompose this forecast into drivers: [forecast question]. Identify the major variables, which ones are most uncertain, which ones matter most, and what indicators I should monitor.

Check calibration

Here are my past forecasts and outcomes: [list]. Analyze whether I seem overconfident, underconfident, vague, or poorly calibrated. Suggest a practice plan to improve.

 

FAQ

Do I need statistics to forecast?

Basic statistics helps, but you can begin with clear questions, base rates, explicit probabilities, and regular review.

Why use numbers instead of words like likely or possible?

Because words hide disagreement. One person’s “likely” may mean 55%; another’s may mean 85%. Numbers force clarity.

What if no base rate exists?

Use the closest reference class you can find, then state the limitation. A weak base rate is not perfect, but no base rate at all lets imagination run around unsupervised.

 

Glossary

    • Base rate: How often something occurs in a relevant reference class.
    • Reference class: A group of similar past cases used to estimate probabilities.
    • Calibration: How well confidence levels match actual outcomes over time.
    • Brier score: A scoring method for probabilistic forecasts; lower scores generally mean better forecasts.

 

References and bibliography

These sources are included so readers can go deeper, check the intellectual foundations, and avoid treating this guide like it descended from the clouds on a glowing clipboard.

    1. Philip E. Tetlock and Dan Gardner, Superforecasting: The Art and Science of Prediction. See also Good Judgment Open’s explanation of probabilistic scoring. Good Judgment Open FAQ.
    2. Richards J. Heuer Jr., Psychology of Intelligence Analysis, CIA Center for the Study of Intelligence. CIA PDF.

 

Next: Causal and Decision Analysis

The next page helps you move from prediction to cause and choice. Forecasting asks what may happen. Causal and decision analysis ask why it may happen and what you should do about it.

This is where you stop arguing about symptoms and start mapping mechanisms, trade-offs, options, and consequences. Reality appreciates the effort, though it will not send a thank-you card.