01
Observe
Collect price structure, filings, estimates, news, ownership signals, and the macro backdrop before forming a view.
7 evidence families
Research methodology
A stock conclusion you can inspect, challenge, and monitor
AlphaVue does not treat a ticker and a prompt as research. It builds an evidence trail, assigns competing analytical roles, records uncertainty, and defines what would change the thesis.
Evidence-led, risk-aware research
Research workflow
Evidence-led, risk-aware research01
Observe
7 evidence families
02
Debate
Opposing cases retained
03
Decide
4-part decision frame
04
Monitor
Event-driven review
Research workflow
One thesis, four controlled stages
Each stage has a distinct job and leaves an auditable output for the next.
02
Debate
Bull, bear, valuation, and risk roles challenge assumptions, source quality, and alternative explanations.
Opposing cases retained
03
Decide
Reconcile the evidence into a directional stance, time horizon, confidence level, and explicit downside case.
4-part decision frame
04
Monitor
Track catalysts, earnings, price breaks, and new disclosures against predeclared thesis-change conditions.
Event-driven review
Evidence coverage
Seven lenses, reconciled rather than blended
Coverage shows the target share of reports in which each evidence family is checked when relevant and available.
Fundamentals
100%Revenue quality, margins, cash generation, balance-sheet resilience, and operating drivers.
Valuation
95%Multiples, cash-flow assumptions, peer context, and the expectations already embedded in price.
Market structure
90%Trend, volatility, volume, liquidity, and support or resistance relevant to timing.
Earnings and estimates
95%Reported results, guidance, revisions, surprise history, and forward consensus dispersion.
News and catalysts
90%Material company events, product cycles, regulation, litigation, and dated catalysts.
Sentiment and positioning
80%Analyst revisions, institutional positioning, short interest, and crowding indicators.
Macro and industry
85%Rates, currencies, commodities, demand cycles, competition, and sector-specific constraints.
Research output
A decision frame, not a naked rating
Directional stance
BUY / HOLD / SELL
A research classification, never a standalone instruction to trade.
Confidence score
0-100
How strongly the available evidence supports the stated thesis.
Risk level
Low-High
The assessed downside profile, independent of directional conviction.
Review condition
Trigger
A measurable event or threshold that requires the thesis to be revisited.
Confidence and risk
Two axes that answer different questions
High / Lower
High confidence, lower risk
Evidence is consistent and key downside channels appear contained; monitoring remains required.
High / Higher
High confidence, higher risk
The thesis is well supported, but volatility, leverage, valuation, or event exposure can still produce severe loss.
Low / Lower
Low confidence, lower risk
Downside appears limited, yet evidence is sparse or conflicting; patience may be more useful than a strong call.
Low / Higher
Low confidence, higher risk
Uncertainty and downside reinforce each other; the report should emphasize avoidance, sizing restraint, and verification.
Known limitations
Where human verification still matters
- Public and third-party data may be delayed, incomplete, revised, or incorrectly mapped.
- AI agents can misread evidence, overgeneralize historical patterns, or produce internally plausible errors.
- Coverage varies by market, security type, reporting regime, language, and availability of reliable sources.
- Confidence is a research-quality assessment, not a calibrated probability of profit or forecast accuracy.
- Reports do not know an individual user's objectives, tax situation, liquidity needs, portfolio, or risk capacity.
Methodology FAQ
Questions investors should ask
Is AlphaVue a registered investment adviser?
No. AlphaVue is an AI-assisted research product, not a registered investment adviser or broker-dealer. Its outputs are research reference, not personalized advice.
What does the confidence score mean?
It summarizes source quality, evidence agreement, and unresolved contradictions. It does not represent the probability that a stock will rise or that a trade will profit.
Why can a high-confidence idea also be high risk?
Confidence concerns support for the thesis; risk concerns possible loss. A well-supported view can still involve leverage, expensive valuation, illiquidity, or binary events.
Does every report use every data source?
No. The system checks evidence families when they are relevant and available. Missing or stale inputs should reduce confidence and be disclosed, not silently imputed.
How are bull and bear views combined?
They are preserved as competing cases. The decision stage weighs source strength, materiality, time horizon, and the conditions under which either case would become dominant.
When is a thesis reviewed?
A review is triggered by scheduled events such as earnings and by material changes in price, guidance, estimates, filings, catalysts, or risk conditions.
Are ratings comparable across all stocks?
They use a common framework, but source depth, liquidity, accounting regimes, and business maturity differ. Comparisons should account for those limits.
Can I trade directly from the report?
The report is a starting point for independent research. Verify material facts and consider your own horizon, sizing, costs, constraints, and professional advice before acting.
Continue researching