Inversion.

Munger’s rule was: tell me where I’m going to die, so I don’t go there. Applied to investing, this is the discipline of asking “how does this thesis kill me?” before asking “how does this thesis make me money?” The cards below are the six failure modes that have, between them, accounted for the great majority of permanent capital losses on otherwise sensible-looking value investments. Each one is an exit door. The job is to know them by name before you walk into the room.

The six failure modes

How a sensible-looking value investment loses you money.

For each: what the failure looks like, a famous example of it happening, and which page on this site is the live guard against it. The order is roughly the order in which they get you — thesis errors first, behavioral errors last.

The moat was an illusion.

The competitive advantage you priced in was already eroding when you bought. The brand, scale, switching cost, or network you believed in could not survive a determined competitor or a structural shift. The thesis depended on durability that wasn’t actually there.

Example Sears. A century of category dominance and a real-estate trove on the books. Both eroded faster than anyone underwriting the thesis was willing to admit.

Guard Reverse DCF: if today’s price implies growth that requires the moat to hold, demand evidence the moat is still widening.

The price wasn’t cheap, just temporarily low.

The earnings in your multiple were peak-cycle, not normalized. The 8× P/E that looked like a gift was really 8× the top of the cycle — which is 25× the cycle average. Cheap-looking ratios on cyclical earnings are the most common value-trap signature.

Example Homebuilders, 2006. Single-digit P/Es on earnings that had been pulled forward by easy credit and would soon collapse to losses for three years running.

Guard Margin of safety: anchor intrinsic value to normalized earnings, not peak earnings. Use 10-year averages when the cycle is in question.

Management destroyed the capital.

Big-ticket acquisitions at peak valuations to flatter EPS growth. Buybacks at the top of the cycle and dilution at the bottom. Adjusted-EPS metrics that exclude every real cost. Each looks fine in a single year. Compounded across a decade, they hollow the business out from the inside.

Example GE under Jeff Immelt. A decade of capital destruction concealed by adjusted-earnings storytelling and serial M&A. The math caught up.

Guard The integrity gate in the Pre-Buy Checklist: a management score below 70 is a hard stop, regardless of price.

The technology made the business obsolete.

The product was great until the substitute arrived. The cash flows you valued were rented from a window in technological history that closed. Categorically, this is the hardest failure to underwrite around, because by the time the substitute is obvious, the price already reflects it; before that, you have to predict.

Example Kodak and Blockbuster. Both had real moats. Both watched their moats become irrelevant inside ten years. Both had analysts defending the thesis right up to the bankruptcy.

Guard Circle of competence. If you cannot name the substitute’s adoption curve five years out, the business is outside the circle — even if you understand it operationally today.

Leverage made small bad news catastrophic.

The business itself was fine. The balance sheet was not. A modest cycle downturn or refinancing window closing turned a recoverable -25% into a permanent -90%. Debt does not change the asset; it changes who owns the asset when the music stops.

Example Highly-leveraged roll-ups in 2008. Equity wiped out at the first refinancing wall, while the underlying operations continued mostly unchanged under new owners.

Guard Net-debt-to-EBITDA < 2.5× for non-financial businesses. Treat “cheap because levered” as a different asset class with different rules.

You sold at the wrong moment.

The thesis was right. The business compounded. You sold at +30% because of anchoring to entry price, and watched it 10× from there. Or you held a structurally broken thesis through a -80% drawdown because you anchored to your own intrinsic-value estimate that had quietly become wrong.

Example Almost every retail investor who held Apple in 2003. Or, in the other direction, almost every retail investor who held a busted thesis in 2008 because it was “cheaper than ever.”

Guard Mr. Market reframing: re-derive intrinsic value periodically, ignore the entry price. Pair with a written sell discipline from the Pre-Buy Checklist.

The pre-mortem

Before you buy, answer these out loud, in writing, by name.

A pre-mortem is the inverted form of the post-mortem. The prompt is: it is three years from now, this position lost money, what happened? If you cannot answer specifically, you are not ready to size the position.

Which of the six failure modes is the most likely way this thesis dies? Name it.
What would have to be true today for that failure mode to be already underway, and how confident are you that it is not?
If management has historically destroyed capital (mode 03), what specifically changed? When? Who is accountable now?
What is the smallest piece of news that, if it crossed your screen tomorrow, would tell you the thesis is broken? Write it down.
If you are wrong, what percentage drawdown have you already pre-committed to absorbing before changing the thesis vs. exiting? Write the number now; not in the drawdown.
What evidence would re-validate the thesis if it temporarily looked broken, so you don’t exit on noise?