Investment thesis · explained
What an investment thesis actually is.
A thesis is a claim about a stock that can be proven wrong — and the credible version says, up front, exactly what would prove it. That single line is the difference between research you can grade and a story you can only believe.
What an investment thesis is
Strip away the jargon and a thesis is four things. A claim — what you expect the stock to do, and why. Its pillars — the few things that have to hold for the claim to be right. Its catalysts — the dated events that will test it. And the one trigger that kills it — the condition under which you accept you were wrong.
Most of what passes for a thesis is only the first part: a confident claim with the reasoning left implicit and no condition under which it could fail. That reads well and tells you nothing. A claim you can act on names the pillars it rests on, so you know what to watch, and the dated catalysts that put those pillars to the test.
Why a thesis has to be falsifiable
A claim is falsifiable when there is a specific condition that would prove it wrong. If nothing could weaken it — if every outcome can be read as confirmation — then the claim is too vague to be useful, because it can never be checked against what actually happens.
Falsifiability is what turns an opinion into something gradeable. Once a thesis carries the exact level or event that would break it, the outcome is no longer a matter of interpretation: the trigger either fired or it didn't. That is the whole basis of a public, scored record — every claim resolves as played-out or invalidated, against the kill condition it published in advance.
An invalidation trigger is not a stop-loss
These get conflated constantly, and the distinction matters. A stop-loss is a price-protection tactic: a level at which you exit an open position to cap a loss, regardless of why the price got there. It is about your capital and your position.
An invalidation trigger is set on the reasoning, not the position. It is the event or level that means the argument was wrong — published in advance, independent of whether anyone holds the name. A stop protects what you own; a trigger tests whether the claim was ever right. You can be stopped out of a thesis that was correct, and you can hold one whose trigger has quietly fired. They answer different questions.
How to evaluate a thesis
Four questions separate a thesis you can critique from one you can only agree or disagree with. They work on anyone's research, including orbyd's.
Is the claim specific?
A thesis that names what has to be true — and why — can be tested. One that gestures at a story can't.
See a dossier dissected →Is there a stated kill condition?
The single event or level that would prove the claim wrong, written before the outcome. No kill condition, no falsifiability.
How we keep it honest →Are both sides argued?
A bull case and a bear case built with the same rigour. A one-sided pitch hides the evidence that would break it.
Browse the dossiers →Is it dated?
When was the claim made, and what did the prior read say? An undated thesis can be quietly rewritten after the fact.
Read the dated journal →Seen in the corpus
orbyd ships 597 dossiers, each with one explicit invalidation trigger published before the outcome is known. When a trigger fires, the thesis is marked broken on the record. A few that resolved that way:
-
Published trigger: A weekly close below $16 loses the post-catalyst breakout shelf (6/18 gap close $16.81) and unwinds the June re-rate; secondary breaks: the AMT-130 Huntington's BLA slips out of the Q3 2026 window or…
-
Published trigger: A weekly close below $7.49 (the 52-week low) confirms capitulation, removes the last shelf, and opens untested single-digit lows; secondarily, FAA TIA flight testing slipping past Q4-2026 or the…
-
Published trigger: A weekly close below $100 confirms the post-print re-rate has fully unwound and opens the gap toward the early-2026 base in the mid-$80s; a theme flip to SATURATED or further clustered PT cuts on the…
This is the concept that every orbyd dossier is built from — a single falsifiable claim with a published kill condition. To see how a frontier model assembles one from news, filings and earnings transcripts, read what AI equity research is, then watch a live one get taken apart part by part.
Common questions
- What is an investment thesis?
- An investment thesis is a falsifiable claim about a stock: what has to be true for the position to work, why, and the specific event or level that would prove it wrong. The credible version names that kill condition in advance, so the claim can be graded later as played-out or invalidated rather than quietly forgotten.
- What is an invalidation trigger?
- An invalidation trigger is the pre-committed condition that, if it fires, means the thesis was wrong. It is published with the claim, before the outcome is known, so there is nothing to rationalise later. The thesis either survived its trigger or it didn't — and the result goes on the record either way.
- How is an invalidation trigger different from a stop-loss?
- A stop-loss is a price-protection tactic about an open position — a level at which you exit to cap a loss. An invalidation trigger is set on the reasoning, not the position: the event or level that means the argument itself was wrong, whether or not anyone holds the name. One protects capital; the other tests whether the claim was ever right.
- How do you critique an investment thesis?
- Ask four questions. Is the claim specific enough to be tested? Is there a kill condition stated in advance? Are both the bull and bear case argued with equal rigour? And is the claim dated, so it can't be rewritten after the outcome? A thesis that fails any of these can't be scored, which means it can't really be checked.
- What makes a thesis falsifiable?
- A thesis is falsifiable when there is a specific, stated condition that would prove it wrong. If nothing could weaken the claim — if every outcome can be read as confirmation — it is unfalsifiable by construction and cannot be graded. Falsifiability is what lets a claim be scored against reality instead of defended forever.