Visit Website →
← Back to blog

Build a Traceable 3 Point Short Thesis From SEC Filings

September 15, 2026
Build a Traceable 3 Point Short Thesis From SEC Filings

A short thesis from filings is a testable hypothesis, not a conclusion, and it becomes tradeable only when at least one independent operational fact corroborates what the disclosures suggest. The minimum evidence standard is simple: the filing signal identifies the anomaly, and a source outside the filing (a former employee, a trade dataset, a site visit) confirms the operational reality behind it. What follows is the stepwise process for getting from a 10-K footnote to a position you can defend.

TL;DR:

  • Valid short theses require at least one independent operational fact confirmed outside the filings, such as trade data or site visits, to verify the disclosed anomaly.
  • The core claim should be supported by two specific filing signals, like rising receivables days or related-party revenue, documented with precise citations.
  • Conducting primary research through interviews, supplier checks, and trade data verification is crucial to test if filings reflect actual operations and avoid false positives.
  • Industry-adjusted financial metrics, especially accruals and gross-margin divergence, help identify operational discrepancies not evident from raw ratios.
  • Publication should be based on a verified chain connecting filings, corroboration, and documented discrepancies, with periodic updates if new evidence emerges.
Lacunaindex
Test Claims Against Corporate Delivery
Lacuna Index uses public records to compare corporate narratives with actual delivery through forensic reports, benchmarks, and execution scores.
Explore Lacuna Index

Table of Contents

Building a Concise Short Thesis From SEC Filings

A workable investment thesis from filings rarely needs more than a core claim and two supporting points. Analysts who pad a report with a dozen disparate observations usually end up defending the weakest one in public. The discipline is to isolate one falsifiable claim, then find the filing items that either support or destroy it.

Start with the sources that produce the highest density of shortable assertions:

  • Form 10-K MD&A section: management's own explanation of results, where optimistic framing often contradicts the numbers two pages later.
  • Footnotes: revenue recognition policy changes, contingent liabilities, and segment reporting shifts frequently hide the real story.
  • Related-party transaction disclosures: look for revenue or cost flows to entities controlled by insiders.
  • Proxy statements: board composition, compensation structure, and auditor tenure reveal governance weaknesses that correlate with disclosure opacity.
  • Form 8-K filings: sudden executive departures, auditor changes, or covenant amendments are often the earliest public tremor before a filing-based thesis becomes obvious to the market.

Once you've scanned those, build the thesis in three layers:

  1. Core claim: state the single sentence that would move the stock if proven (for example, "reported organic growth is substantially inflated by channel stuffing").
  2. Supporting filing signal one: a specific, citable data point (rising days sales outstanding relative to a prior-year quarter, disclosed in the 10-Q).
  3. Supporting filing signal two: a second, independent filing item (a footnote disclosing extended customer payment terms, or a related-party receivable that grew alongside revenue).

Document every claim with the exact filing, section, and fiscal period it comes from. A forensic corporate analysis that can't be traced back to a specific page number is an opinion, not a thesis. Keep a running citation log, because the moment you brief a colleague or investor, you'll be asked exactly where each number came from.

How Do You Test a Filing-Based Short Thesis?

Filings tell you what a company wants you to believe is happening. Primary research tells you what's actually happening on the ground, and the gap between those two things is where a short thesis either holds up or falls apart. The most actionable short reports pair document analysis with investigative fieldwork, because filings alone can produce a defensible paper thesis that collapses the moment someone checks a warehouse, a customer, or a customs record, as RosettiStarr's review of short seller research notes.

Run through this evidence-gathering sequence:

  1. Blind expert calls: commission network interviews with industry specialists without revealing your position or the company name, so answers aren't shaped by your framing.
  2. Former-employee interviews: ask about internal metrics, not opinions. Someone who ran the shipping dock knows more about real volume than someone in corporate communications.
  3. Supplier and customer checks: cross-reference disclosed customer concentration against what suppliers and distributors will actually confirm.
  4. Site visits: a facility that looks understaffed or underutilized relative to disclosed capacity is hard evidence, not speculation.
  5. Trade and customs data: import/export records, satellite imagery of shipping activity, and job-posting trends can corroborate or contradict growth claims independently of anything the company files.

Draft your discussion guides so questions sound like generic industry due diligence rather than a targeted investigation. Asking "how has volume trended this year" gets you further than "have you noticed the company inflating shipment numbers."

Rank your evidence by reliability: direct documentary proof (a signed contract, a customs manifest) outranks a single interview, which outranks a pattern inferred from job postings alone. Muddy Waters Research's Ensign report used site visits and interviews to substantiate a license-renting allegation, precisely the kind of corroboration that turns a filings hunch into a publishable claim.

Evidence types ranked by verification reliability

Pro Tip: Budget two to four weeks and low five figures for a standard verification engagement covering ten to fifteen interviews and one or two site visits. Rushing this stage is how theses go public with holes, which a competent short seller finds on day one.

Scoring Targets With Industry-Adjusted Financial Metrics

Raw financial ratios lie by omission. A gross margin that looks fine in isolation can be an outlier once measured against the right peer group, which is why industry-adjusted metrics are commonly computed by subtracting the annual industry median from a firm's raw figures using SIC/NAICS classification codes. Subtracting the median instead of comparing to a hand-picked competitor set removes a common source of confirmation bias.

Watch these signals in particular:

  • Abnormal accruals: a widening gap between accrual-based earnings and cash flow.
  • Gross-margin divergence versus peers: margins that move opposite the sector median without a disclosed reason.
  • Free cash flow versus EPS gaps: earnings growth unsupported by cash generation.
  • Unusual related-party revenue: a growing share of top-line revenue tied to insider-controlled counterparties.

Academic research on short seller decisions finds that accounting fundamentals like accruals, sales growth, and margin trends predict short interest and forward returns, which supports building a simple weighted rubric: score each signal 0 to 3 based on deviation from the industry median, then weight accrual anomalies and related-party revenue most heavily, since those two categories tend to reflect deliberate structuring rather than cyclical noise.

Once your filing-based score flags a candidate, check market corroboration as a secondary filter. Elevated short interest, unusual options positioning, or activity on platforms tracking prediction market signals suggests other researchers have independently reached a similar conclusion, though it should never substitute for your own primary work.

Why Filings Alone Produce False Positives

A company can be fully compliant with every disclosure requirement and still be operationally hollow. Legal-form compliance measures whether the paperwork is filed correctly, not whether the business does what management claims, and that distinction is where over-reliance on filings turns into a costly mistake. Treating a filing-only pattern as confirmed rather than as a hypothesis is the single most common failure mode in short research.

Watch for these recurring false-positive traps:

  • License-renting arrangements that make regulatory capacity look owned when it's leased from a shell entity.
  • Phantom revenue booked ahead of actual service delivery, especially in healthcare and subscription models.
  • Overstated inventory valuations that inflate current assets without a corresponding cash trail.

Set a stop-loss rule before you size the position, and pause the thesis entirely if primary research contradicts more than one filing signal. Selection bias is the quiet killer here: cherry-picking the filing items that confirm your view while ignoring contradicting footnotes produces a thesis that reads well and trades badly.

Pro Tip: Write down the single fact that would make you abandon the thesis before you publish or size up. If you can't name it, you don't have a thesis yet, you have a bias.

How Lacuna Turns Filing Anomalies Into a Traceable Thesis

A forensic analytics platform can apply an evidence-first logic at scale by mining earnings calls, SEC filings, and proxy statements to quantify the gap between what a company promises and what its own disclosures show it delivered. Sector benchmarks establish the industry-adjusted baseline, and execution and narrative-delivery scores rank which anomalies across a sector are large enough to justify further work.

A simplified sketch of the workflow:

  • A filing shows margin expansion inconsistent with the disclosed cost structure.
  • Corroboration steps confirm or reject the anomaly through independent checks: supplier confirmation, trade data, or a site visit.
  • A documented discrepancy metric quantifies the size of the gap once corroborated.
  • A confidence decision rule determines whether the finding supports a research note, a forensic report, or no action at all.

That chain, filing citation to independent corroboration to documented discrepancy to decision rule, is what separates an audit-traceable finding from an opinion.

The value of a forensic finding isn't the anomaly itself. It's whether the chain from disclosure to independent fact can survive someone else trying to break it.

When presenting findings, separate fact (what the filing states), inference (what the pattern suggests), and decision (what action follows) into distinct sections. Collapsing those three into one paragraph is how defensible research gets mistaken for advocacy.

Sizing, Timing, and the Ethics of Publishing a Short Thesis

Size up only after independent corroboration, not after a compelling filing pattern alone. A thesis resting on filings plus one confirmed operational fact deserves modest conviction; two or three independent confirmations across different evidence types justify real size.

Publication carries an obligation to separate fact from inference explicitly, and to date-stamp claims that depend on conditions that can change, a management response, a subsequent 8-K, an activist letter. Short theses evolve after publication more often than analysts admit, and the honest move is to update the record rather than let a stale claim stand uncorrected. Treat every report as a living document until the position closes, and fold new forensic evidence into portfolio review rather than into a one-time verdict.

— Glen

Where to Verify Filing-Based Claims

Cross-check every claim against primary sources: SEC EDGAR for filings and XBRL data, U.S. Census or customs trade datasets for shipment volumes, state regulatory enforcement databases for licensing history, and the academic literature on short-seller signal accuracy for methodology grounding.

Sources

FAQ

What Is a Short Thesis?

A short thesis is a documented hypothesis that a stock is overvalued relative to its true financial or operational condition, built on specific evidence and structured so it can be proven false.

What Does a Good Short Thesis Example Look Like?

A strong example states one core claim (inflated organic growth), backs it with two independent filing signals (rising receivables days, related-party revenue growth), and confirms at least one signal through primary research such as a supplier interview.

What Is a Three-Point Thesis Statement Example for a Short Report?

A typical structure reads: the core claim (margins are unsustainable), supporting signal one (gross margin diverges from the industry median), and supporting signal two (accrual growth outpaces cash flow), each tied to a specific filing citation.

How Long Does It Take to Verify a Filing-Based Short Thesis?

A standard verification engagement covering interviews and one or two site visits typically runs two to four weeks, though complex operational checks involving trade data or international sites can take longer.

What Makes a Filing Signal Unreliable on Its Own?

A filing signal is unreliable alone when it hasn't been checked against an independent operational source, since companies can be fully compliant on paper while masking a very different operating reality, as seen in cases involving license-renting arrangements.