This InsightMeter guide teaches a short, ruthless habit: before you promote a “smart money” story, try to kill it. Social posts and dashboards love confirmation—new 13F lines, Form 4 buys, a forecaster’s hot streak. Your reader job in twenty minutes is narrower. Write one claim in a sentence a teammate can check, list what would disprove it, and run those tests against primary filings and fixed definitions. If the claim survives, you still do not have a buy tip; you have a claim that survived today’s falsification attempt. If it dies, you saved yourself from recycling theater.
Start a timer. Open a blank note. Do not open a brokerage ticket. Write: “Claim: [who] [did what] in [security] as of [date or quarter], which I interpret as [one meaning], because [one evidence type].” If you cannot fill those brackets, you do not have a thesis—you have a mood. Moods cannot be falsified; claims can.
Falsification is not cynicism. It is the cheapest way to improve research hygiene when filings are lagged, incomplete, and easy to screenshot out of context. A twenty-minute pass will not exhaust every counterargument. It will catch the failure modes that dominate retail “smart money” threads: dollar-value illusions, planned sales narrated as panic, officers and 10% holders treated as one hive mind, and forecast scoreboards with mismatched horizons.
Soft product note: normalized tables can speed lookup, but accession numbers and primary documents still win disputes. Treat any secondary screen as a pointer back to EDGAR, not as the final court.
Vague claims cannot fail. “Smart money loves WidgetCo” is unfalsifiable theater. Prefer operational language: “Manager North’s 13F share count in WidgetCo rose from 1.2M to 2.0M between the 30 Jun and 30 Sep report dates, and I treat that as intentional accumulation rather than a corporate-action artifact.” Now you know what to check: share counts, report dates, amendments, and corporate actions.
Separate observation from interpretation. Observation: “Form 4 on [date] shows code P for 5,000 shares by the CFO.” Interpretation: “the CFO is expressing bullish conviction.” Only the interpretation needs falsifiers. The observation needs verification (accession, amendment, footnotes). Mixing the two is how people argue about feelings while citing filings.
Name the evidence family in the claim: 13F residue, Form 4 change, Schedule 13D/G stake, forecast track record, or a mashup. Mashups are allowed—but each leg gets its own falsifier list. A beautiful mashup with one broken leg is still a broken thesis.
Killer A — 13F share count flat (or down) while the headline screams “buying.” Dollar value can rise solely because price rose. If your claim requires accumulation and shares are flat after checking the same report period, the same share class, and amendments, the accumulation claim is falsified. Prefer share-count deltas over market-value headlines every time.
Killer B — Planned Rule 10b5-1 sales narrated as discretionary bearish conviction. If the Form 4 footnote (or checkbox language your workflow captures) ties the sale to a trading plan, your “insider suddenly hates the stock” story weakens sharply. Planned activity can still be economically large; it is usually weak evidence of a same-day epiphany. Capture plan language before you moralize the trade date.
Killer C — Conflicting Form 4 roles treated as one signal. A cluster that mixes a CEO open-market purchase, a director’s routine award (code A), and a 10% holder’s fund rebalancing is not “insiders agree.” Roles, codes, and economic interests differ. If your thesis needs aligned conviction and the roles conflict or the codes are mostly mechanical, the cluster thesis is falsified or must be rewritten narrowly.
Killer D — Forecast definition mismatch. A “80% accurate” claim dies if the scoreboard used a different horizon, benchmark, hit rule, or sample than the one you are implying for the next call. If the marketed window was cherry-picked or survivors-only, the track-record leg of a smart-money mashup is not evidence—it is advertising.
Add claim-specific killers when needed: 13F lag (quarter-end residue filed later), amendments after your screenshot, recycled social threads with stale as-of dates, or Schedule 13G passive language when someone claims activism. The four killers above cover most twenty-minute failures.
Minutes 0–3: freeze the claim sentence and evidence family. Minutes 3–8: pull the primary document(s)—manager 13F information table, Form 4 with footnotes, or forecast definition card—and save accession or equivalent citation. Minutes 8–15: test the relevant killers with share counts, codes/roles/plans, and scoring definitions. Minutes 15–20: write a one-line verdict: survives / dies / survives only if rewritten as [narrower claim].
Do not spend the twenty minutes searching for more confirming screenshots. Confirmation is infinite on the internet. Falsifiers are finite and usually sitting in the same filing the headline ignored.
If time expires with ambiguous footnotes, mark “unresolved—do not promote.” Ambiguity is a valid educational outcome. Promoting unresolved claims is how recycled threads spread.
Fiction A — “North is loading WidgetCo.” Claim written with share counts. Check: prior quarter 1.8M shares, current 1.8M shares; market value up because price rose. Verdict: accumulation claim dies. Rewrite allowed: “North still holds 1.8M shares as of [as-of]; dollar value rose with price.” That rewrite is boring and accurate.
Fiction B — “CFO dumping is a red flag.” Form 4 shows code S, large size, scary chart annotation. Footnote: sale pursuant to Rule 10b5-1 plan adopted months earlier. Verdict: same-day panic thesis dies or must be rewritten as “planned sale executed; discretionary signal weak.” Size still belongs in the note; motive theater does not.
Fiction C — “Insider cluster buy.” Three Form 4s in a week. One code P by CFO; one code A director grant; one code M/F option exercise and tax withholding that nets little open-market buying. Verdict: conviction-cluster thesis dies. Narrow rewrite: “one open-market purchase by CFO; other lines mechanical.”
Fiction D — “Forecaster X is smart money; follow the new long.” Scoreboard shows high hit rate on 5-day direction calls vs stock absolute return, but the pitch implies 6-month alpha vs sector. Verdict: definition mismatch falsifies the transfer of the scoreboard to the new claim until horizons and benchmarks match.
Copy this checklist. Blank boxes mean the claim is not ready for a teammate—or a stranger on the internet.
Checklists are educational process tools. They do not create profitable trades.
Most failures are social, not intellectual. People protect a thesis they already posted.
Correcting these habits improves research quality even when you never use a commercial dashboard.
Passing a twenty-minute falsification test does not mean the thesis is true, complete, or economically important. Filings omit shorts, many derivatives, intra-quarter trading, and private context. A claim can be consistent with public residue and still be a bad investment idea.
Falsification reduces avoidable errors; it does not produce guaranteed alpha. Markets can reward bad narratives temporarily and punish careful ones. This guide is education about reading public disclosures, not a system for picking stocks.
Some theses die for paperwork reasons (wrong CIK, stale screenshot) even when a better-stated claim might survive. Rewrite narrowly when the paperwork is the problem—do not resurrect the original dramatic wording.
Nothing here is legal advice or a recommendation to buy or sell any security.
Deep due diligence can take days. The twenty-minute pass exists because most viral smart-money claims fail on definitions, not on obscure industry knowledge. Share count vs dollars, plan vs discretionary, role vs hive mind, horizon vs marketing—these are literacy checks.
If a claim survives twenty minutes, schedule a longer session with multi-quarter context, peer manager comparison, and corporate-action adjustment. Do not confuse “survived the smoke test” with “finished research.”
Teams can institutionalize the habit: no smart-money Slack post without a claim sentence and at least one attempted killer. Culture beats tools.
Prefer: “Accumulation claim falsified: shares flat at 1.8M; value up with price; accession …” Avoid: “Smart money is fake.” The second sentence teaches nihilism; the first teaches method.
When a claim survives, say what would falsify it next—upcoming amendment, next 13F share print, Form 4 plan footnote you could not find, or a scoring definition still missing. Survivors should carry their next test date.
Ban certainty language (“proves,” “guarantees,” “can’t lose”) in notes tied to public filings. Filings support modest verbs: consistent with, inconsistent with, insufficient to decide.
Use the recycled-threads guide when the evidence is a social screenshot with suspicious freshness. Use share-count vs dollar-value when Killer A is in play. Use 10b5-1 and Form 4 codes/footnotes when Killers B and C appear. Use forecast scoring and backtest-vs-live guides when Killer D appears.
Falsification is the meta-habit that tells you which literacy guide to open. Without it, you collect articles; with it, you select tests.
Methodology pages that show worked scoring examples make Killer D faster: you already know to demand horizon and benchmark before trusting a percentage.
Spotting recycled “smart money” social threads · Why share-count changes beat dollar headlines · 10b5-1 plans: what they do and don’t tell you · Backtest vs live forecast scoring pitfalls
Write a claim that can die, list disproofs—flat 13F shares, planned 10b5-1 sales, conflicting Form 4 roles, forecast definition mismatch—and spend twenty minutes trying to kill it. Surviving is not a tip; dying is tuition. Continue with related guides, methodology, and glossary. Nothing here recommends buying or selling any security.