Munger borrows a line from the mathematician Carl Jacobi: "Invert, always invert." Plenty of problems that resist a direct assault fall apart when you flip them. Instead of asking "how do I make this portfolio successful?", ask "how could I guarantee this portfolio is a disaster?" Run that and you get a list. Over-concentrate in one factor. Use leverage that forces you to sell at the bottom. Buy stories with no margin of safety, chase whatever just went up, and never sell a broken thesis. That's your checklist of what to avoid. Munger's own version is "All I want to know is where I'm going to die, so I'll never go there." Inversion works for two reasons. Avoiding stupidity is easier, more reliable and more repeatable than achieving brilliance. And the human mind, left to itself, builds a case for what it already wants and stays blind to the case against.
Advertisement
How do you use inversion on a business?
In business analysis, the master inversion question is "How would I destroy this business?" Put yourself in the shoes of the smartest, best-funded competitor, or a hostile regulator, or a technology that makes the product obsolete, and design the attack. If you can't find a credible way to kill the business, you've found evidence of a moat that's far stronger than any bullish narrative. It's Greenwald's barrier-to-entry test, run from the attacker's side. If you find three easy kill-shots, you've found your real risks, and you found them before the market did. Try it on a great business and a fragile one, and feel the difference. "How would I destroy Visa?" is really hard. You'd have to dislodge a two-sided network, thousands of issuing banks and a global merchant-acceptance base built over decades. "How would I destroy a subscale commodity chemical maker with one plant, one big customer and net debt at 4× EBITDA?" answers itself in a sentence.
What is a premortem, and why does it work?
The pre-mortem is inversion turned into a formal, repeatable technique. It comes from the psychologist Gary Klein. Before you commit, imagine yourself three years from now, looking back at a decision that turned out to be a disaster. Take the failure as given and write the story of how it happened. The psychological trick is subtle, and it's real. Asking "what could go wrong?" invites the mind to defend the plan. Assuming the failure has already happened and asking "what happened?" gives people permission to voice doubts they were suppressing, and research finds it markedly increases the number and specificity of the risks that come out. It's the structured cousin of Fisher's scuttlebutt, and the direct ancestor of the "steelman the bear" step in the Phase 8 deep dive.
Worked example: a premortem on a high-multiple stock
A worked pre-mortem. Say you're about to buy a fast-growing, beloved US consumer-technology company on a high multiple, 40× earnings, with a thesis of "durable 20% growth for a decade." Run the pre-mortem. It's 2029, the position lost 70%. What happened? Force out specific stories instead of vague ones:
- Growth decelerated faster than priced. The 20% was pulled forward by a one-time tailwind (a pandemic, a viral product, a subsidy). The base rate for sustaining 20% for a decade is low, and once growth cracked to 8%, the multiple re-rated from 40× to 18×. That's a ~55% hit from the multiple alone, before any earnings miss.
- The moat was narrower than the story. A better-capitalised platform bundled the product for free, the counter-positioning move from 7 Powers, and switching costs turned out to be low.
- Capital allocation destroyed value. Flush with a high stock price, management did a large debt-funded acquisition at the top, and it was impaired two years later.
- A key-person or governance shock. The visionary founder left, and an accounting restatement revealed aggressive revenue recognition.
What do you do with the failure stories?
Here's the payoff. Each failure mode turns into an action. For #1, run the reverse DCF, check the implied growth against Mauboussin's base rates before buying, and size for the multiple risk. For #2, do the 7 Powers analysis and the counter-positioning work explicitly. For #3, read ten years of capital-allocation history and the proxy. For #4, check the founder lock-ups first, then re-read the revenue-recognition footnote. The pre-mortem didn't tell you not to buy. It told you what to verify and what would change your mind, which goes into the decision journal. It also told you how large a position the real risk supports. From here on, every thesis you write carries a pre-mortem. The app templates it, and your personal checklist makes it a standing line.