Decision methods
Compare options when the scores are uncertain
Give each uncertain score a low, base, and high scenario, recompute a weighted sum, and decide whether to choose now or gather evidence first.
A weighted comparison that ends 3.6 to 3.4 looks like a decision. It is only as firm as its inputs, and some inputs are guesses: a cost no one has quoted, or a fit no one has tried.
This guide keeps the weights fixed and varies the scores. Each uncertain score gets a low, base, and high value. These are scenarios you choose, not statistical confidence intervals, and they show whether a ranking survives what you do not yet know.
Set up the comparison
A fictional team is choosing a meeting space. Option A is a rented room and Option B is a co-working membership. Both criteria are scored from 1 to 5, where higher is better. Affordability is weighted 0.6 and schedule fit 0.4, and the weights sum to 1.0.
Write a reason for every bound. The high value for A's affordability equals its base because the rent has been quoted, while the low value allows for a possible cleaning fee. B's schedule fit has wide bounds because nobody has tried booking rooms at busy times. A score with no real doubt can use the same value for low, base, and high.
| Option and criterion | Weight | Low | Base | High |
|---|---|---|---|---|
| A: affordability | 0.6 | 3 | 4 | 4 |
| A: schedule fit | 0.4 | 2 | 3 | 4 |
| B: affordability | 0.6 | 3 | 3 | 4 |
| B: schedule fit | 0.4 | 3 | 4 | 5 |
Compute three scenarios
Base: A = 0.6 × 4 + 0.4 × 3 = 2.4 + 1.2 = 3.6. B = 0.6 × 3 + 0.4 × 4 = 1.8 + 1.6 = 3.4. A leads by 0.2.
A-favorable (A high, B low): A = 0.6 × 4 + 0.4 × 4 = 2.4 + 1.6 = 4.0. B = 0.6 × 3 + 0.4 × 3 = 1.8 + 1.2 = 3.0. A leads by 1.0.
B-favorable (A low, B high): A = 0.6 × 3 + 0.4 × 2 = 1.8 + 0.8 = 2.6. B = 0.6 × 4 + 0.4 × 5 = 2.4 + 2.0 = 4.4. B leads by 1.8.
The leader changes between the stress scenarios, so the base result is not robust.
Find the score worth checking
For each score, calculate its swing as weight × (high value − low value). The swings are 0.6 for A affordability, 0.8 for A schedule fit, 0.6 for B affordability, and 0.8 for B schedule fit. Together they total 2.8, which equals the distance between A's 1.0 lead and B's 1.8 lead.
The schedule-fit scores have the largest swings, and a one-week trial booking at both spaces is cheap. Suppose the trial pins A's fit at 3 and B's fit at 4. The A-favorable scenario becomes 3.6 against 3.4, and the B-favorable scenario becomes 3.0 against 4.0. The leader still changes. All remaining uncertainty is in affordability, so the next step is to get B's price quote and confirm A's cleaning fee.
Choose or gather evidence
If one option leads in the base and both stress scenarios, that supports choosing it within this model, provided it also meets the hard constraints. If the leader changes, look at the score with the largest swing. Gather evidence when that is quick and affordable. If a deadline prevents it, choose anyway and write down which assumption the choice depends on. Another fallback is to compare worst cases. In the original bounds, A's lowest total is 2.6 and B's is 3.0.
Limitations
The bounds are your judgments, and different people could set them differently. The stress scenarios combine every pessimistic or optimistic score at once, which may be less likely than any single change. A weighted sum assumes criteria trade off evenly across the scale, and the weights themselves can be uncertain, which this check does not address.