| name | inversion |
| description | Backward-from-failure thinking — define what would guarantee failure, then systematically avoid it. |
| group | thinking |
| keywords | ["inversion","backward","failure","avoidance","munger","jacobi","contrarian","risk"] |
| task_strategies | ["spike","investigation"] |
| stream_affinity | ["roadmap","research"] |
| allowed-tools | ["Read","Glob","Grep","AskUserQuestion"] |
| status | experimental |
| source | Carl Jacobi (mathematician); Charlie Munger — Poor Charlie's Almanack; mental model popularized by Farnam Street |
| acquired | 2026-06-15 |
Inversion
Invert the problem. Design failure actively. Then avoid everything you designed.
Carl Jacobi: "Invert, always invert." Charlie Munger applied this to business and investing. Instead of asking "how do I succeed?", ask "how do I guarantee failure?" — then don't do that.
When to use
- When you're stuck on how to achieve something
- When forward planning feels abstract or unconvincing
- As a sanity check on any plan ("what would make this fail?")
- Before a high-stakes decision
- Pairs well with
/brana:pre-mortem (pre-mortem imagines failure; inversion actively designs it)
Step 1 — State the goal
What are you trying to achieve?
AskUserQuestion: "What outcome do you want? One sentence."
Step 2 — Invert the goal
Flip it completely:
-
Goal: "Build a product users love"
-
Inverted: "Build a product users despise and never use"
-
Goal: "Ship on time"
-
Inverted: "Guarantee we miss every deadline"
State the inverted goal explicitly. The more specific the inversion, the more useful the next step.
Step 3 — Design failure actively
Generate everything that would guarantee the inverted goal succeeds. Use these prompts:
To guarantee user abandonment:
- What would make the UX maximally confusing?
- What would make support maximally unhelpful?
- What communication would maximize confusion and broken expectations?
To guarantee technical failure:
- What architecture decisions would ensure the system can't scale?
- What shortcuts would create the worst long-term debt?
- What dependencies would be most fragile?
To guarantee team failure:
- What processes would maximize burnout and conflict?
- What would maximize information silos?
- What incentives would create perverse behavior?
Generate 10–20 specific failure recipes. Don't filter.
Step 4 — Recognize current failures
Honest audit: which failure recipes from Step 3 are already present in your current plan or situation?
For each match:
Failure recipe: [X]
Currently present as: [Y]
Severity: high / medium / low
Step 5 — Build the avoidance strategy
For each high/medium severity match:
- What specific change removes or reduces this failure pattern?
- What early warning signal would indicate it's creeping back?
- What habit or process prevents it systemically?
Step 6 — Output
**Inversion analysis of [goal]**
Designed failure patterns: [N total]
Already present: [N high], [N medium], [N low]
Critical avoidances:
1. [pattern] → [avoidance]
2. [pattern] → [avoidance]
3. [pattern] → [avoidance]
Recommendation: [what to stop/avoid immediately]
Notes
- The value is in generating failure patterns, not just listing obvious risks
- The exercise often reveals failures already present that were invisible from the forward direction
- Inversion and pre-mortem are complementary: inversion designs failure deliberately; pre-mortem imagines it already happened
- Munger: "It's not enough to be smart. You have to not be stupid in the ways that are obvious in retrospect."