MAS-Score Methodology
The Mining Asset Score (MAS-Score) is a transparent 8-factor valuation per mining company. Each sub-score (0-100) has its own input data, deterministic formula and AI-generated explanation with clickable source citations.
The 8 sub-scores at a glance
Weight 25%
Weight 15%
Weight 15%
Weight 10%
Weight 10%
Weight 10%
Weight 10%
Weight 5%
Currently 3 of 8 sub-scores are live (Geology, Capital, Management — together 55% weight coverage). The partial score shown reflects these 55%. The remaining 5 sub-scores follow in v0.2 (Jurisdiction, Stage, Permitting Q3 2026) and v0.3 (Catalysts, Promotional Risk Q4 2026).
Geology
Inputs
- ·Grade (g/t)
- ·Tonnage (Mt)
- ·Deposit type (open-pit / underground)
- ·Reserve classification (proven-and-probable / measured-and-indicated / inferred)
Rationale
Measures intrinsic resource quality. High grade values (above 5 g/t for gold) and proven-and-probable classification weight positively. Inferred resources are treated conservatively.
Example
Lundin Gold TSX:LUG: Grade 8.68 g/t × Tonnage 17.98 Mt, P&P-classified → Score 80/100.
Capital Structure
Inputs
- ·Cash position ($M)
- ·Total debt ($M)
- ·Burn rate ($M per quarter)
- ·Shares outstanding (M)
- ·Recent dilution (%)
Rationale
Assesses financial stability and refinancing risk. Net-cash-positive companies with low burn rate receive top scores. Recent dilution above 5% per year triggers penalties.
Example
Lundin Gold: Cash $412M / Debt $95M / Burn -$28M per quarter (positive = free cash flow generative) → Score 100/100.
Management
Inputs
- ·CEO tenure (years)
- ·Insider holding (%)
- ·Board independence (ratio 0-1)
- ·Past discovery record (high/medium/low)
Rationale
Measures management quality and incentive alignment. Long-tenured CEOs (above 5 years) with high insider holding (above 3%) and a track record of successful discoveries weight positively.
Example
Lundin Gold: CEO Ron Hochstein since 2017 (7 years), Insider 4.2%, Board 0.66 independent → Score 95/100.
Jurisdiction
Inputs
- ·Political stability (index)
- ·Mining code quality
- ·Permitting speed
- ·ESG risk per operations country
Rationale
Quantifies country risk based on the Fraser-Institute Mining Investment Index and similar sources. Tier-1 jurisdictions (Canada, Australia, US Nevada) weight positively; Tier-3 (politically unstable countries) trigger penalties.
Example
Ecuador (Lundin Gold) is Tier-2 due to permitting volatility; Nevada (Barrick) is Tier-1.
Stage
Inputs
- ·Lifecycle stage (Exploration / Resource Definition / Pre-Feasibility / Feasibility / Construction / Production)
Rationale
Assesses risk-reward per lifecycle stage. Production stage has low risk but limited upside; Resource Definition has high risk but 10x potential. Score normalized via risk-adjusted returns.
Example
Lundin Gold is Production (score mid 60-70); a junior explorer with a discovery pipeline could reach a score of 75-85.
Permitting
Inputs
- ·Permitting status per asset
- ·Major permits secured
- ·ESG/social-license status
Rationale
Measures regulatory pipeline risk. Companies with all major permits secured (mining lease, water use, environmental approval) weight positively. Open permitting disputes or indigenous-consent conflicts trigger penalties.
Example
Lundin Gold has all major permits for Fruta del Norte; score depends on M&A pipeline permits.
Catalysts
Inputs
- ·Drill results in pipeline
- ·Permitting decisions
- ·Construction milestones
- ·M&A options
- ·Feasibility study updates
Rationale
Assesses forward-looking events within 12 months. High catalyst density (multiple drill results plus permitting decisions plus M&A) weights positively for re-rating potential.
Example
Junior with 3 active drill programs plus a pending PFS (Score 80+); mature producer without near-term catalysts (Score 40-50).
Promotional Risk
Inputs
- ·Stock promotion history
- ·Insider selling patterns
- ·Compliance track record
- ·Press-release inflation index
Rationale
Inverted score — lower values are better. Measures the tendency toward over-promotion. Companies with aggressive PR but thin filings, frequent insider sales or compliance issues trigger high promotional-risk scores (= low valuation).
Example
Established major with conservative filings (Promotional Risk 10-20 = positive for MAS-Score); penny-stock pump pattern (Promotional Risk 70-90 = negative).
Numbers-Guard — how we block AI hallucinations
AI hallucinations are fatal for investment decisions. A fabricated number in an investment recommendation can cause damages in the millions. We address this via a fail-closed pattern:
- The AI generates an explanation with footnote markers
[^1]. - A deterministic Numbers-Guard parses every number from the explanation.
- Each detected number is validated against the input data structure.
- If the AI fabricates a number that is not in the input → retry with strict prompt: "Use EXCLUSIVELY numbers from the input."
- Second guard trip → deterministic template fallback without fabricated numbers.
In the production run of 2026-05-18, for Newmont (NYSE:NEM) the values 134.1 Moz and $3.03B were correctly identified by the guard from the quotedSources — derived values that were not in the primary input structure. Retry with the strict prompt resolved the issue, without the explanation containing any fabricated content.
Source citations
Every AI explanation includes clickable footnote markers [^1] that scroll to the source citation below the score block. When a citation references a documentId that matches in the documents index, the marker becomes an external link to the SEDAR+/EDGAR PDF — otherwise an in-page anchor to the quotedSource text.
Academic standard, no marketing-speak: every number-based statement has a traceable source. In the UI this is visually rendered as numbered italic citations below each sub-score.
See the MAS-Score in action.
Four companies live indexed. Lundin Gold (TSX:LUG) shows all 3 sub-scores plus 3 source documents (NI 43-101, MD&A, Proxy Circular) with clickable citations.
View Lundin Gold profile →