Promotional Risk in Junior Mining Is Three Numbers, Not a Feeling

Promotional risk in junior mining is usually judged by tone — a press release that reads too excited, a stock that's suddenly everywhere on social media. MAS-Score scores it from three measurable components instead, each traced to a specific SEC filing: how much of a company's press-release volume is substantive rather than hype, what share of insider holdings has been sold in the trailing twelve months, and whether a company's own proxy statement discloses a promotional history for its CEO. None of the three is a read of tone.

The first component: a ratio, not a count

A company that issues 30 press releases a year is not automatically more promotional than one that issues 5 — the question is what fraction of that volume is substantive. Mineralis's event-layer pipeline pulls every 6-K and every 8-K carrying Items 2.02, 7.01, or 8.01 (earnings, regulation-FD disclosures, and other material events) in the trailing 12 months, then classifies each one. A deterministic first tier resolves most releases outright — an item code, an exhibit over roughly 1 MB, or attached XBRL data is enough to call a filing substantive without a model in the loop. What's left goes to a language model constrained to quote the release verbatim, and a second, independent model checks that classification before it counts — a disagreement between the two is a veto, not a tie-break. The resulting substance ratio is combined with a volume floor that only engages for genuinely high-frequency issuers: 5 points of added risk at 12 or more releases a year, 10 at 24 or more, 20 at 40 or more, regardless of how substantive each one is individually.

The second and third: selling and history

Insider selling is read directly from Form 4 filings — the percentage of an insider cohort's reported holdings sold within the trailing 12 months, with no offsetting purchases treated as the clearest case. The penalty is tiered: 8 points at 10% or more sold, 18 at 20% or more, 25 at 30% or more. The third component, ceoPromotionalHistory, is a boolean read from the company's own proxy statement or annual information form rather than from news coverage. The rule leans on a specific piece of securities law: Regulation S-K Item 401(f) legally obligates a U.S. issuer to disclose certain legal or regulatory proceedings involving a named officer, in that officer's own biography. A CEO biography that carries no such disclosure is therefore read as a documented absence — coded false, meaning no known promotional history — rather than left blank because nothing negative happened to be mentioned. A citation-validated extraction feeds a second model that has to independently confirm the same reading; a contradiction resolves to unknown rather than to whichever answer sounds more confident.

Missing data is not scored as clean

Promotional Risk runs on an inverted scale — 100 is worst, 0 is clean — which makes the handling of missing inputs a real methodological choice, not a formality. Foreign private issuers file no Form 4, and Canadian issuers report insider trades through SEDI, which has no public bulk API; for those companies, insider-selling data is currently unavailable at all. Silently dropping that component would make an unmeasured company score better than one that's known to be clean, simply because nothing could be checked. Mineralis instead rescales the components that are known across the remaining assessable weight, and where every known component is already at zero, the sub-score reports zero rather than manufacturing a penalty for a gap in coverage — the same fail-closed instinct behind the Numbers-Guard. In a five-company pilot measured as of September 10, 2026 — Barrick, Newmont, Agnico Eagle, Southern Copper, and Lundin Gold — this left the insider-selling and CEO-history fields null for three of the five (Barrick, Agnico Eagle, and Lundin Gold), purely because their insiders report through SEDI rather than Form 4, not because anything about their conduct is in question[^3].

What the same pilot measured for the two SEC filers

Newmont (NYSE:NEM) furnished 7 press releases under the qualifying item codes in the trailing 12 months to the measurement date, and all 7 classified as substantive — a ratio of 1.0. Six named executives, including its CEO, sold a combined 115,287 shares in the same window with no recorded purchases, against a Form-4-based denominator that the pipeline itself treats as an upper bound, since it counts only the insiders who filed in the window rather than the company's full share count. That put Newmont's insider-selling read at roughly 10.3%. Its CEO history read false: the proxy statement's biography for Natascha Viljoen, President and CEO, carries no Item 401(f) disclosure. Southern Copper (NYSE:SCCO) furnished 8 qualifying releases, all substantive, with an insider-selling read of roughly 3.5% and the same false read on CEO history[^3].

A ratio built for the small end of the market

Those substance ratios of 1.0 are not a coincidence of the pilot cohort — they're close to a structural floor. What a senior producer furnishes to the SEC under Items 2.02, 7.01, and 8.01 is, by construction, close to always material: earnings, financed transactions, executive changes. The ratio has very little room to discriminate at that end of the market. Its purpose is the opposite end: a small-cap issuer with a high volume of releases carrying little disclosed substance is exactly the historical pattern regulators describe. The SEC's Office of Investor Education and Advocacy has warned for years about the mechanics of that pattern in microcap stocks generally — a trading surge with no material news, unsolicited "inside" information, huge return predictions, and paid promotion that isn't disclosed as such[^1]. The British Columbia Securities Commission's 2022 settlement with Bearing Lithium Corp. (then Bearing Resources Ltd.), a TSX-V-listed junior lithium explorer, is a mining-specific instance of the same undisclosed-promotion pattern: between January and February 2017 the company retained a marketing firm that placed an advertorial written to read like independent journalism, plus posts from 19 social-media accounts across Twitter, LinkedIn, Facebook, investFeed, and iHub promoting its Chilean lithium project — none of it disclosed as company-sponsored[^2]. The settlement, $35,000 combined between the company and its CEO, wasn't reached until August 2022 — more than five years after the promotional conduct itself, illustrating how far behind the paper trail an enforcement outcome can land relative to the promotion it addresses.

What this means for coverage

Promotional Risk carries a 5% weight in the overall MAS-Score — smaller than Geology (25%) or Capital and Jurisdiction (15% each), but structurally different from the others in one respect: its inputs depend on which disclosure regime a company reports under. A U.S.-listed junior filing Form 4 is, today, more fully measurable on this sub-score than a TSX-V-only name reporting insider trades through SEDI — which is exactly the segment where the Bearing Lithium case shows the underlying risk concentrates. A null reading on this sub-score is not a clean bill of health; it is an honest statement that the evidence isn't assembled yet, and closing that gap for SEDI-reporting issuers is next on the roadmap rather than a problem quietly left unscored.

References

1. U.S. Securities and Exchange Commission, Office of Investor Education and Advocacy, "Updated Investor Alert: Fraudulent Stock Promotions," last updated March 29, 2016. Scope: general warning on microcap and penny-stock promotion schemes across U.S. markets; not mining-specific and not a mining-sector statistic.
2. British Columbia Securities Commission, Settlement Agreement, Bearing Lithium Corp. (formerly Bearing Resources Ltd.) and Jeremy Arthur William Poirier, settled August 4, 2022 (announced August 5, 2022). Scope: a single TSX-V-listed junior lithium explorer regulated in British Columbia, Canada. The underlying promotional conduct (an advertorial and posts from 19 social-media accounts) occurred January–February 2017 — the settlement date is more than five years later and should not be read as the date of the conduct itself.
3. Mineralis internal event-layer measurement, "Event-Layer, Schritt 1: promotionalRisk aus EDGAR-Filings," pilot run dated September 10, 2026. Scope: a five-company pilot cohort (Barrick, Newmont, Agnico Eagle, Southern Copper, Lundin Gold) built to validate the extraction mechanism, not a statement about promotional risk across the covered universe; three of the five companies have no Form-4-based insider data because they report through Canada's SEDI system instead.