Over the past year, Your Science | Research has focused on developing and upgrading its semiautomated stock analysis engine, AuditAlpha™. AuditAlpha™ covers three investment horizons — days to months, six months to two years, and ten years or more — through three corresponding modules: PulseLens, BottleneckLens, and CompounderLens. Powered by hundreds of pages of purpose-built Python code, each module applies a methodology tailored to its horizon, but all three share the same core principles: systematic reliable-data retrieval, transparent decision rules, documented reasoning, and fully auditable investment analysis.
Sample Reports
The three sample reports below illustrate the capabilities of AuditAlpha™, currently in beta, on selected securities. They are not investment advice. Once finalized, AuditAlpha™ will be offered by subscription; in the meantime, we welcome feedback on the reports to guide its further development.
PulseLens™ – Short-Term Investment Quality Engine
PulseLens is an almost fully automated investment analysis framework that retrieves live market data, consensus targets, and corporate and macroeconomic event flags — covering corporate calendars, Federal Reserve, ECB, and FRED releases. It synthesizes these inputs into a structured, multi-layer analytical stack spanning macro regime, fundamentals, technical configuration, volatility, and risk management. Each criterion is evaluated as PASS, FAIL, or CONDITIONAL with full diagnostic traceability. This transparency creates a clear audit trail, ensuring that every verdict rests on a verifiable and reproducible analytical record. Designed for private investors and professional portfolio managers seeking rigorous, bias-free decision support.
BottleneckLens™ — Medium-Term Investment Quality Engine
BottleneckLens is a semiautomated analysis protocol for the months-to-years holding horizon, evaluating a candidate across ten criteria organized in three layers — Moat, Timing, and Risk — each returning a PASS, FAIL, or CONDITIONAL verdict with High|Medium|Low confidence, rolled up into Moat and Timing scores that feed a structured verdict table. Alongside this initial assessment, a dedicated Risk Management suite runs on a weekly cadence: a three-tier Technical Status model weights confirmed institutional selling most heavily, structural price confirmation second, and short-term price action least. A companion Peak Psychology module tracks three independent signals associated with a stock nearing a cycle top — good-news rejection, analyst-estimate dispersion, and options positioning. This core assessment, Technical Status and Peak Psychology, runs weekly. A Daily Briefing analyzes news, disclosures, and sector-wide developments across all tracked tickers — not merely summarizing news flow, but connecting it to each position’s underlying investment thesis — while a separate, monthly Bubble Signal Checklist tracks systemic bubble risk across the broader ecosystem. Every weekly cycle is preserved to a persistent history, so the engine can summarize a position’s trailing trajectory, not just its current snapshot. As with every AuditAlpha™ module, the human analyst reviews the evidence and finalizes the verdict — the engine assembles the case, never decides alone.
CompounderLens™ — Long-Term Investment Quality Engine
Built for the 10-year-plus holding horizon, CompounderLens assesses a candidate in three stages. A Fundamental Durability gate tests seven criteria — earnings power, accounting quality, sustainable returns on capital, cash-flow authenticity, balance-sheet resilience, concentration risk, and capital stewardship — as a strict, all-must-clear gate rather than an averaged score: a single genuine structural failure disqualifies a candidate outright. For each criterion, the engine retrieves the underlying evidence directly from primary sources such as SEC EDGAR, runs a set of purpose-built deterministic tests, and poses the remaining qualitative questions — each one classified as resolving a flagged test, corroborating an already computed result, or complementing it with evidence no test could capture — to the assisting LLM, evidence attached. Every answer is returned with its supporting evidence, its reasoning, and a confidence level, determining whether its criterion is satisfied — with all seven required to clear the Fundamental Durability gate — and may inform the subsequent Competitive Advantage and Reinvestment Logic assessment and the scenario-based Total Shareholder Return model (Bear|Base|Bull, across multiple holding periods). As throughout AuditAlpha™, the analyst reviews every LLM judgment and finalizes the consensus verdict — nothing is decided by the model alone, and every step remains traceable in a full audit trail.
