What is an investment thesis invalidation condition?

An invalidation condition is the answer to a single question, written down before you take the position: "What specific, observable event would tell me my reasoning is wrong?"

It is not a stop-loss. A stop-loss is a risk-management device tied to price; it says "I will not lose more than X dollars on this position." An invalidation condition is tied to the thesis; it says "if Y happens, the story I told myself to justify this position is no longer true, regardless of where the price is."

The two often disagree, and that disagreement is where the value lives. A stop-loss can trip on noise while your thesis remains perfectly intact — you take a small loss for nothing. Worse, a stop-loss can fail to trip while the thesis has quietly died around you — the position drifts sideways, you hold it for months, and you eventually exit at a larger loss with no idea what happened.

Good invalidation conditions share three properties. First, they are observable: a specific data release, a specific price level held for a specific duration, a specific qualitative shift you can point to in a 10-K. Second, they are pre-committed: you wrote them down before opening the position, so when the trigger fires you can't rationalize it away. Third, they are mechanism-specific: they target the load-bearing link in your reasoning chain, not the conclusion.

Example. Thesis: long ASML because EUV monopoly makes unit demand resilient to a single-quarter capex pullback, and operating leverage on installed-base service revenue protects margins. Bad invalidation: "stock down 20%." Good invalidation: "TSMC and Samsung jointly pause High-NA orders for more than two consecutive quarters" OR "service revenue mix falls below 24% for two prints." Both target the actual mechanism. Either trip means the thesis is dead, even if the stock is up.

Most investors resist writing invalidation conditions because doing so requires them to admit, in advance, that the thesis is contingent on something specific. It's easier to stay vague. But vague theses can't fail — and a thesis that can't fail can't teach you anything either.