Quant staples—MVO and Risk Parity—win by estimating risk well. They are "no‑alpha" frameworks whose real value lies in robust risk matrices (vols, covariances). Their favor often tracks recent performance rather than out‑of‑sample validation.

• Equal‑Weight → assumes equal vol, zero corr, equal Sharpe to be mean‑variance optimal.
• Naïve Risk Parity → inverse‑vol weights; relaxes equal‑vol assumption.
• Full Risk Parity → full covariance; assumes equal Sharpe across assets.

Large platforms favor dense, Gaussian‑like libraries (hundreds of signals) — but this often dilutes conviction. We choose a sparse, Laplace‑like prior: fewer, economically durable drivers where weights matter.

• High‑conviction factor: Innovation, persistent and economically grounded.
• Avoid "averaging out" alpha across 1000s of weak signals.
• Interpretability & accountability for every basis point of risk.

We forecast returns by measuring firms' innate innovation capacity and its translation into revenues, margins, and market share— aligned with endogenous growth economics. Long‑horizon, low‑turnover, benchmark‑agnostic.

• Single‑factor sort on Innovation with robust estimation.
• Quant model → long‑run signal; fundamental lens → near‑term viability.
• Equal‑weight with innovation tilt; strong active share.

• Closet indexing → Correlation ~1, fees overwhelm tiny active bets.
• Benchmark anchoring → Arbitrary universes constrain alpha.
• Crowding → Consensus trades dilute edge.

Our response: High active share, factor‑first selection, no benchmark hugging.

Quant models surface candidates; fundamental review tests causality, moat, management, and near‑term risks.

• From ratios → statistical histories → event‑driven ML features.
• Signals reflect business behavior, not just accounting artifacts.

• Innovation: R&D efficacy → commercial outcomes.
• Adaptive Capacity: capital reallocation speed.
• Quality of Growth: margin & FCF accretion.

Traditional quant models treat capex as a black box. LLMs unlock earnings calls, transcripts, and management commentary to systematically classify capital allocation—something impossible with financial ratios alone.

• Measures effective innovation (outputs & commercial impact), not raw R&D spend.
• Captures outperformance of firms that create new profit pools.
• Implicit diversification vs demographic & secular headwinds.

Our research shows a clear performance hierarchy: Innovation Leaders consistently outperform Innovation Laggards, who in turn outperform non-R&D payers.

We're witnessing the birth of a new quant paradigm. LLMs enable systematic analysis of business archetypes and patterns that previously required elite fundamental analysts with decades of experience.

Based on backtested results and forward-looking estimates.

• Balanced weights; caps on single‑name & sector concentration.
• Covariance‑aware diversification (avoid hidden common factor overload).
• Sector‑agnostic selection → hedge against demographic & secular drags.