The Probative methodology
What we do, and why we hold ourselves to it
Version 2.0 · June 2026 · Free to read and share
Foreword
This is a short paper about how Probative works.
We will not describe our internal tooling, our prompts, our calibrations, our ranking formulas, or the specific thresholds we apply at each stage of our process. That work is the actual moat behind the publication and we keep it to ourselves — the way a chef keeps a recipe, the way a fund keeps its model.
What we will describe is what we believe, what we have committed to, and what you will see from us week to week. The receipts are public. The reasoning is paywalled to Members. The system is ours.
Part I — What we believe
Owning a share of stock means owning a small piece of a real business — its factories, its brand, its contracts, its cash, its debt. The stock's price on any given day is the public market's best guess at what that piece is worth. Sometimes the guess is reasonable. Sometimes it is wildly wrong.
Value investing is the discipline of identifying when the market's guess is meaningfully too low — buying then, waiting until the market's guess catches up, and selling. It is a 90-year-old idea, written down by Benjamin Graham, executed by Warren Buffett for sixty years at Berkshire Hathaway. The playbook is in every public library.
We are in one specific corner of that tradition: buy good businesses that are temporarily cheap. Do not buy bad businesses that are permanently cheap. A "good business" is one whose returns on the capital it invests are durably high, whose moat is observable in the filings, and whose cash flow shows up year after year. "Temporarily cheap" is when the price reflects a problem that the evidence in the company's own filings shows is temporary — a one-time charge, a passing macro fear, a stretch of weak quarters in a cyclical industry — rather than a problem that the evidence shows is permanent.
The whole work is telling the temporary kind of cheap apart from the permanent kind. That is what we do. That is all we do.
Part II — The three commitments
These are the rules we hold ourselves to. They will not change.
1. Filings only.
Every claim we make about a company comes from a specific paragraph in a specific government-filed document — the 10-K, the 10-Q, an 8-K, the proxy statement, the XBRL financial statements. We do not rely on press releases, earnings-call transcripts, analyst opinion, or our own memory of the company.
Why: corporate filings are written under penalty of perjury. Everything else is marketing.
2. Cite the source.
Every key claim in every memo carries a clickable citation that opens at the exact paragraph it came from. A typical Probative memo carries thirty citations. A claim without provenance is not a claim — it is an opinion, and we do not publish opinions as facts.
Why: every reader can audit every claim. You don't have to trust us; you can check.
3. Hold to a rule.
What turns a verdict into a system rather than an opinion is that we decide each name against a rule we wrote down before we saw the company — and we follow the rule whether we like the answer or not. The rule is not the model's opinion, and it is not ours. It is a discipline that exists in a separate place from the analysis. When the rule says BUY, we open the position. When the rule says PASS, we pass — even on a name we like. When the rule's exit criteria fire on a position we hold, we re-think — by the rule, not by feel.
Why: human judgment is good at reading filings and bad at consistent decision-making. Separating the two is what gives us a track record we can stand behind years later.
Part III — What we will never do
A negative list, because it matters.
- We will not trade. Holding period is years, not days. We will not react to a single bad quarter. We will not react to a macro headline. The thesis is judged against the filings, on the filings' clock.
- We will not use leverage. Every position is long, cash-funded, sized at conviction.
- We will not present opinion as fact. Every claim has a citation or it does not appear.
- We will not edit the scoreboard. Every verdict we issue — BUY, WATCH, or PASS — lives on a public page from the day it ships, with realized return next to it. We do not delete the bad ones. We do not re-date the good ones. We do not move losses to private memos.
- We will not hide our holdings. Every position we open appears in the public ledger the day it opens. Every exit appears the day it closes.
- We will not promise outperformance. Past performance does not predict future results. The receipts are receipts; they are not a guarantee.
Part IV — What you will see
We owe a publication a clean view of what it is. Probative's surface is small on purpose.
One letter, every Tuesday morning, at 8:00 AM Eastern. One name analyzed in full, with the thesis, the bear case, the trigger that would prove us wrong, and the link to the cited memo where every claim is sourced. Or, when the calendar warrants, a check-in on a name we already hold.
The public scoreboard — every verdict we have ever issued, with realized return versus the Russell Mid-Cap benchmark, updated continuously. No edits. No deletions.
The live ledger (Members) — every position we hold, the date it opened, the entry price, the current price, the return, the distance to fair value. Updated through the trading day.
The cited memos (Members) — the full thesis on every name we hold, every claim hyperlinked to the SEC paragraph that supports it.
Alerts (Members) — when one of our internal exit criteria fires against a position, you find out the same day.
That is the publication. There are no special "founders' tiers," no Discord, no podcast guests, no consulting calls, no model marketplace, no API. There is a letter, a scoreboard, a ledger, and the memos behind them.
A note on our academic foundations
Our work draws on a small canon of academic and practitioner sources. The intellectual lineage is public: Graham and Dodd on intrinsic value, Buffett on quality and patience, Greenwald on the asset-replacement view of value, Piotroski on quality scoring, Altman on distress, Cohen, Malloy and Nguyen on what changes in filings actually mean. Anyone serious about value investing should read all of them. The publication you are reading is one disciplined application of that canon; it is neither the only one nor the inevitable one. We share the lineage; the operational layer is ours.
Coda
Probative exists because we believe the discipline of reading filings closely, citing every claim, and deferring to a written rule is rarer in this business than it should be — and because the only honest way to demonstrate that a system is real is to put its receipts on a public page from day one.
That is the entire pitch.
Cite the source. Hold to the rule.
— The Probative