AInvestor score: What it means
The AInvestor score is a composite measure used internally to rate stocks based on quality and valuation metrics.
What the score represents
A higher AInvestor Score indicates stronger combined fundamentals and market signals relative to peers. The score is intended as a quick comparison tool and a first-pass filter, not a substitute for full due diligence.
How the score is constructed (step-by-step)
- Collect feature set: assemble component metrics for the company such as profitability (ROIC, EBITDA margin), growth (revenue or EPS growth rates), cash generation (FCF yield), leverage (net debt / EBITDA), valuation (P/E, EV/EBITDA relative to peers), and price momentum.
- Sanity filtering: ensure required inputs exist and are numeric; mark missing values and apply sensible defaults or flags. Certain metrics require positive denominators and are excluded when invalid.
- Winsorize / cap outliers: limit extreme metric values (e.g., top/bottom percentiles) to reduce sensitivity to erroneous or one-off values.
- Normalize metrics: convert raw metrics to a common 0–100 scale using percentile or z-score mapping within an appropriate universe (sector or global peer set). For example, higher ROIC maps to a higher normalized sub-score while higher leverage maps to a lower sub-score.
- Apply weights: multiply each normalized sub-score by its predefined weight. Typical weight buckets include Quality (profitability, return metrics), Growth (revenue/EPS growth), Valuation (relative P/E, EV/EBITDA, FCF yield), Momentum (price-based signals), and Financial Health (leverage, liquidity).
- Aggregate: sum the weighted sub-scores and rescale to a 0–100 range. If any critical metrics are missing, the aggregation applies a penalty or adjusts the denominator so the score remains comparable.
- Stability smoothing: optionally blend the current-score with recent historical scores to avoid excessive score volatility from a single data revision.
- Output and flags: return the final numeric score plus component breakdowns (sub-scores, weights), validation flags (e.g., missing inputs), and a success indicator. Downstream consumers use these fields for display, sorting, and filtering.
Interpretation & practical notes
- Scores are relative to the scoring universe used for normalization (sector vs. global can change percentiles).
- A score of 80+ typically indicates strong, across-the-board metrics; 40–60 is neutral; below 30 suggests material weaknesses.
- The component breakdown is as important as the headline score — a high score driven purely by momentum differs from one driven by fundamentals.