A falsifiable investment thesis is one that names specific observable conditions under which the investor would conclude the thesis has failed. Most investment theses fail this standard not because investors are careless but because the failure modes are systematic and feel reasonable in the moment they occur. This guide identifies the five most common ways investment theses avoid falsifiability, with concrete examples and fixes for each.
Why falsifiability failures are systematic, not accidental
Investors do not usually intend to make unfalsifiable theses. The failure comes from the way investment reasoning is typically structured: starting with a conclusion and working backward to support it, rather than starting with observations and testing a hypothesis.
When an investor identifies a company they find attractive, their natural inclination is to build a case for owning it. They gather evidence that supports the case, interview management, read favorable sell-side research, and construct a narrative. At the end of this process, they have a rich set of supporting evidence and a strong conviction. What they often do not have is a specific prediction about the future, a defined time frame, or a clear description of the evidence that would indicate the thesis has failed.
This structure is not irrational. Building a case is how investment analysis works. The problem is that conviction built this way tends to produce theses that are descriptive rather than predictive, theses that monitor the wrong evidence, and theses with no natural exit condition. The five failure modes below are the most common patterns. Each one has a diagnostic sign and a structural fix.
Failure mode 1: The thesis is a description, not a prediction
The most basic falsifiability failure is a thesis that describes present conditions rather than predicting future outcomes. Consider a thesis that reads: "This company has a great product, a loyal customer base, and excellent management that has consistently allocated capital well." Every part of this statement may be true. None of it is falsifiable, because none of it makes a claim about what will happen in the future.
A description cannot be falsified. If the stock underperforms for three years, an investor holding a descriptive thesis can always respond, "The product is still great, customers are still loyal, and management is still excellent. The market simply hasn't recognized the value yet." This is a coherent response to underperformance, but it is also a perfect escape from any accountability to a thesis.
The diagnostic sign is that the thesis can be stated equally well whether the stock has gone up 50% or down 50% in the past year. If the thesis survives unchanged regardless of what the stock has done, it is probably descriptive rather than predictive.
The fix is structural: rewrite the thesis as a prediction with a time-bound expected outcome. "This company's enterprise software segment will grow revenue at 20% or more annually through 2028 as it captures market share from legacy on-premise vendors" is a prediction. It names a metric, a magnitude, a time frame, and a mechanism. Each element can be monitored and each can be falsified by observable evidence.
Failure mode 2: Moving goalposts after evidence arrives
The second failure mode is the single most common pattern among experienced investors, precisely because it is the one that most resembles legitimate analytical updating. The investor changes the falsifier or the time frame in direct response to evidence that would otherwise have triggered a reassessment. This is not updating a thesis based on new information; it is avoiding a verdict by redefining what would constitute one.
An example: an investor writes a catalyst thesis with the falsifier "if the FDA approval does not arrive by Q4 2026, the time-sensitive catalyst thesis is invalidated." Q4 2026 passes without approval. The investor's response is not to reassess the position but to extend the time frame to Q2 2027, citing a regulatory pipeline update as evidence that approval is still forthcoming. Q2 2027 passes. The time frame extends again.
The investor believes they are rationally updating in response to new regulatory information. From the outside, the pattern looks like an unfalsifiable thesis dressed in updating language. The key diagnostic is whether the change to the falsifier or time frame is driven by genuinely new information about the underlying business or regulatory environment, or whether it is driven by the investor's desire to preserve the position.
The fix is procedural: write falsifiers before entering the position and date-stamp them. When the time horizon or a falsifier needs to change, the investor should document the original falsifier, the date of the change, the reason for the change, and what specific new information justified the change. This creates an audit trail that makes the distinction between legitimate updating and goalpost moving visible and reviewable.
Failure mode 3: Evidence threshold is vague
Many investment theses name the right metrics to monitor but fail to specify how much movement in those metrics would be material. "If growth slows significantly," "if margins deteriorate meaningfully," "if the competitive position weakens" are not evidence thresholds. They are placeholders for decisions that have not yet been made.
The problem with vague thresholds is not that they cannot eventually trigger an exit. The problem is that they require the investor to make the threshold judgment in real time, under conditions where bias is strongest. When margins have declined by 150 basis points over three quarters, is that "meaningful deterioration"? The investor who wrote "meaningful deterioration" as their falsifier must now judge this in the context of an existing position, existing conviction, and existing sunk cost. The answer they arrive at will be systematically biased toward the conclusion that 150 basis points is not quite material enough.
A real threshold names a specific metric, a specific direction, and a specific magnitude. "If gross margin falls below 45% in any two consecutive quarters before the end of fiscal 2027" is specific. It removes the in-context judgment about whether a given data point counts. Two investors reading this independently will agree on whether it has been triggered, regardless of their position.
The fix is to write thresholds with enough specificity that the triggering question has a yes-or-no answer. Metrics should be named from the company's actual financial reporting, so there is no ambiguity about where to find the data. Direction and magnitude should be stated as specific numbers, not relative adjectives. Time frames should be stated as calendar dates or fiscal periods, not as abstract windows.
Failure mode 4: No time frame defined
"This will eventually be recognized by the market." The word "eventually" is not a time frame. It makes the thesis unfalsifiable by definition, because any horizon, however long, is consistent with the thesis remaining intact. An investor holding an "eventually" thesis can hold through five years of underperformance and remain entirely rational within the terms of the thesis.
This failure mode is particularly common among investors who think of themselves as long-term fundamental investors. The long-term orientation is legitimate; the refusal to name a time frame is not. Even a genuine long-term thesis can be written with a defined window. A 10-year compounding thesis with a 2035 review date is falsifiable in a way that an "eventually" thesis is not.
The diagnostic sign is that the investor cannot say what year, or even what decade, they expect the thesis to play out. If asked "when would you conclude this thesis has failed?" the investor's answer is "when the business fundamentals deteriorate" without reference to time. This conflates the falsifier (a business event) with the absence of a time horizon (when the falsifier must trigger to constitute a failure).
The fix is to name the time frame when the thesis is written. The time frame does not have to be precise to the quarter, but it must be precise enough to be actionable. "By the end of 2029" is workable. "Eventually" is not. If the investor genuinely cannot name a time frame, that is diagnostic: the thesis type may be unclear, or the thesis may be a narrative that has not yet been converted into a testable prediction.
Failure mode 5: Tautological break condition
"I will exit when I realize I was wrong about the story." This is the most sophisticated-sounding form of an unfalsifiable thesis, because it appears to include a break condition. It does not. A break condition that depends on the investor's realization of error is a break condition that can only be triggered after the stock has already moved decisively in one direction.
If the stock doubles, the investor will not realize they were wrong about the story; they will feel that the story has been validated. If the stock halves, the investor's realization of error will be indistinguishable from loss aversion and post-hoc rationalization. Either way, the "break condition" is determined by price performance rather than by observable business metrics.
The same problem appears in softer forms: "I will exit if the investment case deteriorates." "I will revisit if the fundamentals stop supporting the thesis." These sound like monitoring commitments but they all reduce to "I will exit if I change my mind," which has no observable trigger independent of the investor's existing beliefs and existing position.
A valid break condition must be based on observable business metrics that can be measured independently of price movement and independently of the investor's existing conviction. The condition should be capable of being triggered even while the stock price is rising. If a break condition could only fire when the stock has already fallen substantially, it is likely functioning as a loss-recognition rule rather than a thesis-integrity rule.
How to audit your existing thesis for these failures
A structured falsifiability audit takes five questions and applies them to the written thesis:
First: does the thesis make a specific prediction about a future outcome, or does it describe current conditions? A descriptive thesis fails the first test.
Second: have the falsifiers or time frames changed since the thesis was written? If so, was each change documented with the date and the specific new information that justified it? Undocumented changes are likely goalpost movements.
Third: could each evidence threshold trigger from reported financial data, without additional investor judgment about magnitude? A threshold that requires the investor to judge whether a given data point is "significant" is too vague.
Fourth: is there a specific date or period at which the thesis will be evaluated? A thesis with no time horizon cannot be falsified by underperformance.
Fifth: could each break condition be triggered independently of price movement? Read each break condition and ask whether it could fire while the stock price was unchanged or rising. If not, it is likely price-based rather than fundamentals-based.
A thesis that passes all five questions has a reasonable claim to being falsifiable. A thesis that fails one or more questions has a structural problem that is worth addressing before it becomes the source of a bad exit decision.
Frequently asked questions
What is the most common falsifiability failure in investment theses?
The most common falsifiability failure is the thesis that is a description rather than a prediction. Statements like "this company has a great product and strong management" describe current conditions but name no specific future outcome that could be verified or refuted. Without a named outcome and a time frame, there is nothing to falsify.
How is moving the goalposts different from legitimately updating a thesis?
Legitimate thesis updating occurs when genuinely new information changes the fundamental analysis, such as a new competitor entering the market or a regulatory change. Goalpost moving occurs when the investor changes the falsifier or the time frame specifically in response to evidence that would otherwise have triggered a reassessment. The test is whether the change is driven by new external information or by the investor's desire to avoid concluding the thesis has failed.
Why is a price decline not a valid falsifier for a fundamental thesis?
A fundamental thesis makes claims about business performance, competitive position, or valuation gaps. A price decline is consistent with a fundamental thesis being correct, particularly when the overall market sells off, liquidity is constrained, or the time horizon has not yet been reached. Using price as the falsifier collapses a fundamental thesis into a momentum observation, which is a different investment approach entirely.
What is a tautological break condition?
A tautological break condition is one that can only be triggered by price movement, making the thesis self-confirming. "I will exit when I realize I was wrong about the story" is tautological because the investor's realization will inevitably be shaped by price. If the stock has risen substantially, the story feels confirmed. If it has fallen substantially, the story feels wrong. This reduces the thesis to a price chart interpretation rather than a business evaluation.
How do I write an evidence threshold that is specific enough?
A specific evidence threshold names three things: the exact metric being monitored, the direction and magnitude of movement that would be meaningful, and the time window over which that movement is measured. Instead of "if growth slows significantly," write "if revenue growth falls below 10% for two consecutive quarters before the end of 2027." The threshold should be concrete enough that two investors reading it independently would reach the same conclusion about whether it had been triggered.