When is a simple what-if check enough for a low-stakes choice?

Use a simple check only for a bounded decision

A simple what-if check asks: if one important assumption were reasonably lower or higher, would the same small next step still make sense? It is enough only when the choice is genuinely low-cost and low-risk, the action is easy to pause or revise, and no hard boundary changes across the range. Stability supports a bounded next move; it does not prove the forecast correct.

HM Treasury's 2026 Green Book says analytical effort should be proportionate to a proposal's scale, cost, complexity, and risk. It specifically allows simple what-if questions for low-cost, low-risk proposals, while pointing to more advanced techniques for proposals with higher cost, risk, complexity, or irreversibility. This page adapts that proportionality principle to personal reflection; it does not turn an everyday check into formal appraisal.

Run the LIGHT check

LIGHT is a Forever Free Compass framework, not an external research finding. Write one line for each field before treating a quick what-if as adequate.

1. Low stakes: name the maximum plausible cost, time, disruption, and downside, and confirm the step can be paused or revised.
2. Identify one assumption: choose the uncertain input most likely to change what you do next, not merely the easiest number to vary.
3. Guardrails intact: confirm that permission, consent, duties, safety conditions, affordability limits, and other hard constraints hold across the range.
4. Hold the rest steady: change only that assumption so you can see whether it actually controls the result.
5. Test a plausible range: write a defensible low and high value, then check whether either one changes the next move.

Choose one decision-critical assumption

Start with a literal statement such as, “I assume this trial will take two to four hours,” or, “I assume I can reverse this commitment after one week.” Prefer an assumption that connects directly to the decision. If changing it would not affect the next move, it is not the useful what-if for this choice.

The Green Book describes sensitivity analysis as testing how changes in key assumptions affect results and recommends stress-testing assumptions that are uncertain or susceptible to change. It also describes a simple range as a reasonable low and high around an estimate. For an informal check, use a range you can explain from available facts or experience; do not invent probabilities or extreme endpoints to manufacture stability.

Match the evidence effort to the question

Use only enough evidence to make the range and boundary credible for this next step. A calendar, quoted price, written policy, short observation, or direct confirmation may be more useful than a large collection of loosely related information. Record where the range came from and what remains unknown.

CDC evaluation guidance says evidence and indicators should align with the question being answered. It says method choice depends on resources, ethics, scope, validity, and reliability, and it advises balancing the amount collected with burden. The CDC guidance is for program evaluation, not personal decisions; the limited adaptation here is to make evidence collection question-led and proportionate rather than exhaustive.

Read the result without overstating it

If the same bounded next step remains reasonable at both ends of the range and every guardrail still holds, the simple check is enough for that step. Record the assumption, range, stable next move, and review point. You have shown robustness to one stated uncertainty, not eliminated uncertainty or established that the outcome will occur.

If the preferred next move flips, do not average the endpoints or choose the result you like. Gather one piece of evidence that would narrow the range, reduce the size of the action, preserve both options, or use a fuller comparison. If several uncertain inputs interact, change together, or cannot be bounded credibly, a one-variable check is no longer informative enough.

Raise the standard when consequences or duties are material

A quick what-if is not enough when a choice could materially affect health, legal rights, regulated money, employment, confidentiality, immediate safety, consent, or another person's welfare. It is also not enough for an expensive or hard-to-reverse commitment, a decision with meaningful tail risks, or a choice governed by a formal review process. Use the required process and qualified guidance instead.

UK SAGE guidance is explicitly for emergency advice, so it is not evidence that an emergency method should be copied into an everyday choice. Its boundary principle is useful here: uncertainty and limitations should be made explicit, and no single strand of advice or evidence should be the sole basis for consequential decisions. LIGHT therefore stops at the edge of low-stakes reflection rather than pretending one check can carry a high-stakes decision.