Decision Sensitivity Analyser
A weighted decision matrix gives an answer; what it does not give is any sense of how much that answer depends on weights someone chose in a meeting. This scores the options, then sweeps each criterion's weight to find where the winner changes — and rescores the whole decision at that weight rather than estimating it from the margin. The criteria that can flip it are the ones worth arguing about.
What this generator does
Multiplies each option's scores by the criteria weights to get totals, names the winner, and then varies each criterion's weight across its whole range to find the point at which a different option comes top.
How to use this tool
- List your criteria with a weight for each.
- List your options with a score against every criterion.
- Read the winner and its margin.
- Read the tipping points — those are the weights that actually matter.
Understanding the controls
- Criteria
- One per line as a name and a weight. Weights are relative, so what matters is their ratio rather than their scale.
- Options
- One per line as a name followed by a score for each criterion, in the same order. Any consistent scale works — nought to ten is usual.
Common use cases
- Testing how robust a weighted decision really is
- Finding which criterion's weight is doing the deciding
- Focusing a discussion on the weights that change the answer
- Checking whether a close decision is close for a real reason
- Presenting a decision with its sensitivity rather than just its result
How this generator works
Each option's total is the sum of its scores times the criteria weights. To find a tipping point the tool holds every other weight fixed, sweeps the one in question from zero upwards, and rescores the entire decision at each step — so a reported flip is one that genuinely happens rather than one extrapolated from the margin. Before display the totals are recomputed, the winner is confirmed to have the highest, and every tipping point is re-tested at the weight it names.
Randomness and fairness
Nothing here is random. A set of weights and scores gives exactly one result and one set of tipping points.
For how randomness is produced across the whole site, see how Generate Random works.
Limitations and good to know
- One weight is varied at a time, so this misses combinations where two weights move together.
- Weights are swept over whole numbers from zero to a hundred, so a tipping point between two integers is reported at the first whole weight that flips it.
- Scores are taken as given; the sensitivity of the answer to a mis-scored option is a separate question.
- A robust answer is not necessarily a right one — it only means the weights are not what is deciding it.
Privacy and your data
The analysis runs in your browser and your criteria, options and scores are never transmitted, stored or included in analytics.
Related generators
- Decision Matrix GeneratorScore options against weighted criteria — with cost and risk counting the right way round, and a check on how solid the answer is.
- Weighted Yes/No Decision GeneratorGet a yes/no answer with an adjustable bias, e.g. 70% likely to say yes.
- Fair Division GeneratorSplits indivisible items between people who value them differently, then asks each of them whether they would rather have someone else's share.
- Seat Apportionment GeneratorTurns vote totals into whole seats by D'Hondt, Sainte-Laguë or largest remainder, with every allocation summing exactly and compared against its quota.