The Method

The Q/V/M Framework

Every stock in our universe receives three independent scores, each on a 0–100 scale. These map to the three dimensions that academic research and practitioner experience have shown to drive long-term returns:

Each pillar is built from multiple underlying factors, equally weighted within their group. The factor weights are configurable but must always sum to 1.0 within each pillar.

Quality Factors

Quality captures franchise strength and financial health. We look at three groups of metrics:

Franchise Quality

Return on Invested Capital (ROIC) and Return on Equity (ROE). High, stable returns on capital signal a durable moat.

Piotroski F-Score

A 9-point checklist covering profitability, leverage changes, and operating efficiency. Scores of 7+ indicate strong financial health.

Leverage

Debt-to-equity and interest coverage ratios. Lower leverage means less fragility in downturns.

Value Factors

Value compares market price to fundamental anchors. We use a broad set of ratios to avoid single-metric traps:

PE & PB

Price-to-Earnings and Price-to-Book. Classic valuation multiples compared against sector medians.

Price-to-Free-Cash-Flow

PFCF rewards companies generating real cash, not just accounting profits.

Price-to-Sales

PS is useful for cyclical or early-stage businesses where earnings are volatile.

EV/EBITDA

Enterprise value over operating earnings normalises for capital structure differences.

Dividend Yield

A tangible return to shareholders that also signals management confidence in cash flows.

Momentum Factors

Momentum captures the market’s own signal. Stocks that have been rising tend to keep rising, and improving earnings often precede further gains.

Price Momentum

3-month, 6-month, and 12-month relative price performance versus the universe. We exclude the most recent month to avoid short-term reversal noise.

Earnings Momentum

Analyst estimate revisions and earnings surprise trends. Upward revisions are a strong leading indicator.

How Scoring Works

Raw fundamentals are transformed into comparable scores through a multi-step process designed to be fair across sectors and robust to missing data.

Sector-Relative Percentiles

Every metric is ranked within its sector, so a cheap utility doesn’t automatically beat an expensive tech stock. Apples are compared to apples.

Thin-Sector Blending

Sectors with fewer than 20 stocks get a weighted blend of sector and universe percentiles, preventing small-sample distortions.

Missing Data Penalty

Missing values receive a score of 25 (below median) rather than being ignored. No free passes for incomplete filings.

StockRank Re-Percentiling

The final score is the re-percentiled average of Quality + Value + Momentum. Re-percentiling spreads scores across the full 0–100 range, counteracting the Central Limit Theorem’s tendency to bunch averaged percentiles toward 50.

StockRank: The Composite Score

The three pillar scores are combined into a single StockRank from 0 to 100. Each pillar contributes equally by default, though the weights are configurable. A StockRank above the BUY threshold generates a buy signal; below the SELL threshold triggers a review.

Decile tracking validates that higher-ranked stocks genuinely outperform lower-ranked stocks over rolling periods, giving the model a measurable trust score.

AI Conviction Layer

Quantitative scores alone can miss breaking news: fraud allegations, regulatory action, or transformational deals. Our AI news filter adds a qualitative overlay.

For every stock that passes the quantitative screen, Google Gemini analyses recent news from DuckDuckGo, Finviz, SEC EDGAR, and NewsAPI. Each stock receives one of three verdicts:

Stocks with missing AI verdicts default to FLAG as a safety measure, ensuring nothing enters the portfolio without a clean bill of health.

Macro Regime Detection

Portfolio sizing and rebalancing cadence adapt to the prevailing market regime. Four macro indicators are combined into a composite regime signal:

VIX

Implied volatility index. Elevated VIX signals market stress.

Yield Curve

10Y minus 2Y Treasury spread. Inversion has historically preceded recessions.

Credit Spreads

Investment-grade corporate bond spreads over Treasuries. Widening spreads indicate tightening financial conditions.

Fed Policy

Federal funds rate trajectory and forward guidance tone.

The composite maps to three regimes: RISK_ON (monthly rebalancing, full position sizing), NEUTRAL (biweekly rebalancing, standard sizing), and RISK_OFF (weekly rebalancing, reduced sizing, tighter stops).

Sell Discipline

Knowing when to sell is at least as important as knowing what to buy. Three rule-based triggers enforce disciplined exits:

Stop-Loss

An absolute floor below the entry price. If a position drops past this level, it is sold automatically to cap downside.

Trailing Stop

A dynamic stop that ratchets up with the high-water mark. Locks in gains while giving winners room to run.

Thesis Review Trigger

If a stock's StockRank deteriorates below the SELL threshold or receives a REMOVE verdict from the AI filter, the position is flagged for immediate review.

Position sizing uses a modified Kelly criterion with conservative fractional sizing, sector caps, and maximum position limits to prevent concentration risk.

Academic Foundations

Every factor in our scoring system traces back to peer-reviewed research. The table below summarises the key papers and their findings.