Verified against ChatGPT · 2026-08-13
Build a best/base/worst case scenario model around one specific financial decision
Structures a three-case scenario analysis around a stated decision, forcing each case to change one clearly named variable rather than producing three vaguely different vibes labeled optimistic, likely, and pessimistic.
The prompt
Ready to copy — highlighted parts are example details you can swap.
Build a three-case scenario model — best, base, and worst — around the specific decision below. This is a structuring exercise to organize the requester's own assumptions clearly, not a prediction of what will actually happen. DECISION BEING EVALUATED Whether to open a second retail location given uncertain foot traffic in the new area. BASE CASE ASSUMPTIONS Expected monthly foot traffic of 4,000, conversion rate of 8%, average sale $34, fixed monthly costs for the new location $9,200. KEY VARIABLE THAT DRIVES THE OUTCOME Monthly foot traffic — could range from 2,500 (worst case, slow area) to 6,000 (best case, if a nearby anchor store draws more people than expected). WHAT'S BEING MEASURED Monthly net profit or loss for the new location. BUILD RULES Identify the one or two variables that most drive the outcome (the key variable given, plus any other the base assumptions imply matters) and change only those between the three cases — do not let unrelated assumptions drift between cases for no stated reason, since a scenario analysis is only useful if the reader can see exactly what's different between cases and why. For each case, state the specific value each key variable takes, not just a label like "optimistic" — a scenario without a stated number for its key driver isn't a scenario, it's a mood. Show the resulting output metric for all three cases using the same calculation method, so the three results are genuinely comparable rather than each computed a slightly different way. State explicitly which case you're treating as most likely to occur, if the requester's assumptions imply one, or say that likelihood wasn't specified and all three should be weighed without an assumed probability. Do not present the base case as a forecast or promise — it is one plausible path built from stated assumptions, not a prediction endorsed with confidence. OUTPUT FORMAT 1. Table: case | key variable value(s) | resulting output metric | one-line rationale for why this variable value represents this case. 2. A sensitivity note: how much the output metric changes per unit change in the key variable, so the requester can see how sensitive the whole model is to that one number being off. 3. What this model does not account for — factors outside the stated assumptions that could still change the outcome.
Customize
Optional — swap in your own details for the highlighted parts above.
Why this works
The rule that only the stated key variables should change between cases, with everything else held constant, is what prevents scenario analysis from collapsing into three vaguely different narratives — a common failure mode where an "optimistic" case quietly assumes better foot traffic AND lower costs AND a shorter ramp-up period all at once, which makes it impossible to tell afterward which assumption actually drove the better outcome. Requiring a specific numeric value for the key variable in every case, not a label like "optimistic" or "pessimistic" on its own, forces GPT-5.1 past the point where it could otherwise produce three qualitatively different-sounding paragraphs without any of them being pinned to a number the requester could actually check their real-world outcome against later. The sensitivity note — how much the output moves per unit change in the key variable — exists because the real value of a three-case model isn't the three numbers themselves, it's understanding how fragile the base case is to being wrong about one assumption; a model that shows profit swinging wildly with a small change in foot traffic tells the requester this decision is far riskier than one that stays roughly profitable across a wide range, and that's a different piece of information than the three headline numbers alone convey. Refusing to present the base case as a forecast or promise matters because a three-case model built from the requester's own assumptions is only as good as those assumptions — the model has no independent way to verify that 4,000 monthly visitors is realistic for this specific location, so treating the base case as a confident prediction rather than one plausible path would overstate what the exercise actually establishes.
What you get back
Worst case: 2,500 monthly visitors -> net loss of roughly $1,900/month. Base case: 4,000 visitors -> net profit of roughly $1,700/month. Best case: 6,000 visitors -> net profit of roughly $6,900/month. Sensitivity: each additional 500 monthly visitors adds approximately $1,360 to monthly profit at the stated conversion rate and average sale — the model is highly sensitive to foot-traffic assumptions, which weren't independently verified.
Verified against
ChatGPT GPT-5.1 · 2026-08-13
Changelog
- 2026-08-13 — Initial publish, verified against ChatGPT GPT-5.1.
Need this built into your business?
If a prompt isn't enough — what Scult builds, built and maintained for you — that's Scult's day job.
EXPLORE WHAT SCULT BUILDS
