How Quantitative Signals Drive AI Portfolio Construction

Quantitative signals are the foundation of systematic investing. Rather than relying on narrative-driven stock picks, a signal-based approach converts measurable data into a composite score that ranks every stock in the universe on the same scale. Early Signal builds these signals across three distinct timeframes, each capturing different market dynamics.

Three Timeframes, One Composite

Markets move on different clocks. Earnings surprises drive price action over hours, analyst revisions reshape consensus over weeks, and capital expenditure trends play out over quarters. By separating signals into daily, weekly, and quarterly components, the engine avoids conflating short-term noise with structural shifts.

Daily Signal

The daily composite blends price momentum (30%), options flow (20%), news sentiment (20%), earnings surprise (20%), and insider trading activity (10%). This captures the fast-moving information that drives intraday and next-day price action.

Weekly Signal

The weekly composite emphasizes analyst estimate revisions (30%), sector rotation patterns (25%), supply chain leading indicators (25%), and key technical levels (20%). These factors reflect the medium-term consensus shifts that institutional investors act on.

Quarterly Signal

The quarterly composite focuses on fundamental shifts in revenue acceleration and margin trends (35%), valuation multiple resets (25%), capital expenditure trend analysis (20%), and market share changes (20%). These slow-moving drivers determine the structural winners and losers over multi-quarter horizons.

Layer-Aware Blending

Not every layer in the AI supply chain responds to the same timeframe equally. Upstream commodity layers (raw materials, semiconductor manufacturing) are driven more by quarterly fundamentals, while application-layer companies respond faster to daily sentiment shifts. The engine assigns layer-specific weights to each timeframe component so that the composite reflects each stock's actual information environment. See the full weight tables on the methodology page.

Conviction and Disagreement

The composite score maps to a conviction level: HIGH (≥0.75), MEDIUM (≥0.50), or LOW (≥0.25). When the three timeframes disagree in direction, a 30% disagreement penalty dampens the composite. This ensures the engine only takes strong positions when multiple timeframes confirm the same thesis.

Conviction directly drives portfolio construction. High-conviction names enter the core sleeve, medium goes to satellite, and low-conviction daily signals feed the tactical allocation. Explore the live signal heatmap to see how signals distribute across the AI ecosystem today.

Try It Yourself

Free accounts get a preview of the signal heatmap with the top 10 ranked stocks. Operator accounts unlock the full signal table across all timeframes and tickers.