Finding direction
How can I compare options when my outcome estimates are uncertain?
Compare uncertain options in two passes: first remove any that fail a hard constraint, then compare the rest against the same criteria using ranges or plain-language bands rather than single confident forecasts. Identify the assumption most likely to change the ranking and run a what-if check: if that estimate becomes less favorable, does your preferred option change? Prefer an option that remains acceptable across plausible outcomes—not automatically the one with the highest best case. If a small reversible test can reduce a ranking-changing uncertainty, test before making a larger commitment.
Set the comparison before estimating outcomes
Start with the decision you actually own, its time horizon, and a short list of realistic options, including the status quo when it is real. Then write the criteria each option must be compared against. University of Minnesota Extension's criteria-grid method likewise begins by listing criteria and comparing every option across the same columns.
Mark any true requirement separately. An option that breaches a non-negotiable limit involving consent, safety, rights, duties, affordability, or access should not regain viability through a strong score elsewhere. For the remaining preferences and benefits, define what weak, acceptable, and strong performance would mean before favoring an option.
Use the RANGE comparison
RANGE is a Forever Free Compass framework, not an external research finding. Complete the five fields in order and keep the underlying notes visible.
- Requirements: remove or redesign options that fail a genuine hard constraint; record why the requirement applies.
- Anchors: choose the same few decision criteria for every viable option and define what each criterion means in this decision.
- Numbers as ranges: replace unsupported point forecasts with an evidence-based range or a plain-language band, and name what makes the estimate uncertain.
- Give the ranking a stress test: vary one decision-critical estimate at a time and record the change that would make another option preferable.
- Experiment or exit: if a small reversible action can reduce a ranking-changing uncertainty, run it; otherwise choose the option that remains acceptable across the most credible conditions or pause for qualified evidence.
Show uncertainty instead of hiding it in a score
GovS 010 says decisions should assess options against defined criteria and consider uncertainty holistically. It also says sources and impacts of uncertainty should be identified at a level proportionate to the decision, with results communicated without unwarranted confidence in one option. That standard governs government analysis, not personal decisions, but it supports a useful interpretation: the uncertainty belongs beside the estimate, not buried after the total.
Use a numerical range only when you have a reasonable basis for both ends. Otherwise use explicit bands such as weak, acceptable, and strong, followed by the evidence and missing information. Do not assign invented percentages merely because a spreadsheet accepts them.
| Comparison field | What to record | Warning sign |
|---|---|---|
| Criterion | The same definition for every option | The definition changes to favor one option |
| Estimate | A supported range or explicit qualitative band | A precise number with no defensible basis |
| Evidence | Source, relevance, and important limitation | A vivid anecdote treated as a forecast |
| Switch condition | The change that would alter your preference | No imaginable result is allowed to change the ranking |
Stress-test the preferred option
The 2026 Green Book describes sensitivity analysis as testing how changes in key assumptions affect results. It also says simple what-if questions may be proportionate for low-cost, low-risk proposals. For a personal comparison, ask one plain question at a time: if the most important estimate moved toward its credible downside, would the preferred option change?
Record the switch condition. For example: “Option A remains preferable unless the weekly time requirement regularly exceeds my available boundary.” A close or easily reversed ranking is a signal to gather better evidence, shrink the commitment, or preserve flexibility. A ranking that survives reasonable changes is more robust, though it is still not guaranteed to produce the preferred outcome.
Do not let one total make the decision
A criteria grid can make comparisons inspectable, and the University of Minnesota Extension describes a weighting variation when choices are close. But the Green Book cautions that an initial ranking or single summary metric is not sufficient on its own for a balanced judgment; risks, uncertainties, and factors not captured in that metric still matter.
Treat any total as a conversation starter. Read across the row: which criteria drive the result, which estimates are least reliable, what is missing, and whose interests are affected? If small, arbitrary weight changes repeatedly reverse the winner, report a close comparison instead of manufacturing a decisive score.
Choose the next evidence, not impossible certainty
Government Analysis Function guidance says information about uncertainty that is critical to sensible use should be prominent, and enough context should be available for users to judge whether an estimate fits its intended purpose. Applied here, your comparison should make the decision-critical unknown easy to see and connect it to a next action.
For a low-stakes choice, that action might be a sample task, a schedule simulation, a written quote, or a precise question to an informed person. When health treatment, legal rights, regulated finances, immediate safety, employment duties, confidentiality, or material effects on another person are involved, do not rely on RANGE alone. Obtain qualified input, required consent, and reliable domain evidence before acting.
Original contribution
The Forever Free Compass take
Forever Free Compass uses RANGE as an original comparison framework: Requirements, Anchors, Numbers as ranges, Give the ranking a stress test, and Experiment or exit. RANGE separates hard constraints from preferences, compares every viable option against the same decision criteria, exposes the assumptions behind uncertain estimates, and asks what reasonable change would switch the preferred option. It does not calculate a scientifically validated score, supply missing probabilities, or make a fragile numerical total decisive. Its purpose is to reveal whether the choice is robust enough for a bounded next step or whether one uncertainty needs better evidence first.
Sources
- The Green Book (2026) — HM Treasury. Balanced judgment beyond one summary metric; uncertainty and sensitivity analysis; switching values; simple what-if questions for low-cost, low-risk proposals.
- Government functional standard GovS 010: Analysis — UK Government Analysis Function. Defined decision criteria; holistic and proportionate treatment of uncertainty; assumptions, options, and scenarios; balanced communication without unwarranted confidence.
- Decision making techniques — criteria grid — University of Minnesota Extension. Listing decision criteria; comparing every option across the same criteria; documenting the comparison; optional weighting when choices are close.
- Communicating quality, uncertainty and change — UK Government Analysis Function. Clear, prominent communication of uncertainty, limitations, and fitness for intended use so people can apply appropriate weight to estimates.
This is educational reflection, not medical or mental-health care. If distress is persistent, severe, or involves immediate safety, contact a qualified professional or local emergency service.