Surfaces the criterion you're actually optimising for, tests each option against being wrong, and names the information that would settle it — or tells you it's already a coin flip.
I keep going back and forth between two options. OPTION A: {{OPTION_A}} OPTION B: {{OPTION_B}} WHY I'M STUCK: {{STUCK}} **1. What am I actually optimising for?** From how I described these, infer the criterion I am really using — not the one I would say out loud. Name it in one sentence. If I appear to be optimising for two things that conflict, say so; that is usually the whole problem. **2. The reversibility test.** For each option: if it turns out wrong in a year, what does undoing it cost? Cheap-to-reverse decisions deserve less agonising than I am giving them. **3. What I lose either way.** Name the specific thing each option forecloses. Do not soften this — the fear of loss is what has me stuck, and seeing it written down usually shrinks it. **4. The one fact that would settle this.** Something I could find out this week. If there genuinely isn't one, say so. **5. Your call.** Pick one and give the single strongest reason. If the honest answer is that these are close enough that the choice matters less than committing, say exactly that. Rules: - Do not produce a weighted scoring table. I will game the weights to get the answer I already want. - Do not tell me to trust my gut. - Do not hedge in section 5. Pick one. - Work only from what I told you. Ask nothing.
A prompt is a skill with the serial numbers filed off. Install it once and your agent reaches for it on its own.
Paste this into ChatGPT, Claude Code, Codex, Cursor, or whatever you use:
Install the decide-between-two-options skill globally from https://promptsbuddy.com/prompts/decide-between-two-options/skill.md.claude/skills/decide-between-two-options/SKILL.md
---
name: decide-between-two-options
description: "Surfaces the criterion you're actually optimising for, tests each option against being wrong, and names the information that would settle it — or tells you it's already a coin flip. Use when the user asks for help with strategy tasks like decisions, thinking, choices."
license: CC-BY-4.0
metadata:
source: https://promptsbuddy.com/prompts/decide-between-two-options
author: "Rony"
version: "1.0"
---
# Decide between two options you keep going back and forth on
Surfaces the criterion you're actually optimising for, tests each option against being wrong, and names the information that would settle it — or tells you it's already a coin flip.
## Inputs to collect first
Ask the user for anything below that they have not already given you. Do not
invent values for these.
- `OPTION_A` — The first option, and what draws you to it. (e.g. Take the bigger job — more money, more scope, a manager I don't know.)
- `OPTION_B` — The second, and what draws you to that. (e.g. Stay — work I like, people I trust, no obvious next step.)
- `STUCK` — Why you haven't decided. This is usually the most revealing field. (e.g. Every time I pick one I immediately think about what I'd lose from the other.)
## Instructions
I keep going back and forth between two options.
OPTION A: {{OPTION_A}}
OPTION B: {{OPTION_B}}
WHY I'M STUCK: {{STUCK}}
**1. What am I actually optimising for?** From how I described these, infer
the criterion I am really using — not the one I would say out loud. Name it in
one sentence. If I appear to be optimising for two things that conflict, say so;
that is usually the whole problem.
**2. The reversibility test.** For each option: if it turns out wrong in a
year, what does undoing it cost? Cheap-to-reverse decisions deserve less
agonising than I am giving them.
**3. What I lose either way.** Name the specific thing each option forecloses.
Do not soften this — the fear of loss is what has me stuck, and seeing it
written down usually shrinks it.
**4. The one fact that would settle this.** Something I could find out this
week. If there genuinely isn't one, say so.
**5. Your call.** Pick one and give the single strongest reason. If the honest
answer is that these are close enough that the choice matters less than
committing, say exactly that.
Rules:
- Do not produce a weighted scoring table. I will game the weights to get the
answer I already want.
- Do not tell me to trust my gut.
- Do not hedge in section 5. Pick one.
- Work only from what I told you. Ask nothing.
## Notes from people who use this
- Section 1 ends most of these. 'You're optimising for not feeling foolish' lands harder than any pros and cons list.
- The reversibility test is the practical one — most decisions people agonise over are cheap to undo, and most they rush are not.
- If it says the options are close enough that committing matters more, believe it. That's a real finding, not a dodge.
---
Decide between two options you keep going back and forth on · by Rony · v1.0
From PromptsBuddy — https://promptsbuddy.com/prompts/decide-between-two-options
Licensed CC BY 4.0.
Every prompt is also available at /prompts/decide-between-two-options/skill.md — see all install options.
Weighted decision matrices feel rigorous and are theatre: you adjust the weights until the total agrees with the answer you already wanted, then feel justified. Banning the table forces the actual questions — what am I optimising for, what does being wrong cost — which are harder and more useful.
The most practical section. People spend weeks on decisions they could undo in a month, and minutes on ones they cannot undo at all. Asking what reversal costs reallocates the worry to where it belongs.
Works out what you actually want from it, what the other person is likely protecting, and the first two sentences — the part everyone rehearses badly.
DineshHands the model a plan and asks it to write the post-mortem for a failure that hasn't happened yet. Surfaces the risks your team is too invested to name out loud.
Rony