Portfolio Credibility: Scoring Your Book Against an Evidence-Based Baseline

Why Credibility Scoring Catches What Conviction Doesn't

Conviction is self-reported — you believe in the name. Credibility is externally-anchored — does the evidence support the weight you're giving it? The two often diverge, and the diverge is where mistakes live. You have 8% in a name you "love," but its credibility score is 35 because revenue decel is real, insiders are selling, institutional ownership is dropping, and technicals are topping. That's the exact pattern that leaks alpha. Credibility scoring exists to surface the gap early.

What the Portfolio Credibility Card Shows

The Portfolio Credibility card scores every holding on a 0-100 composite built from five evidence families: fundamentals (revenue + margin + FCF trend quality), valuation (percentile vs sector + reverse-DCF implied growth reasonableness), technicals (relative strength, trend structure, breadth), flow (institutional ownership delta, insider MSPR, options positioning), and regime fit (whether the name historically performs in the current macro regime). Each family is weighted equally by default; premium users can tune weights. A portfolio-weighted-average credibility score summarizes the whole book. Positions ranked by credibility let you see immediately which holdings are strong on evidence and which aren't.

Reading the Scores

Three behaviors tend to improve portfolios over time. First, trim positions with credibility < 40 and size > 3% — over-sized, under-supported positions are the category that hurts most in a drawdown; shrinking them costs nothing in opportunity if the evidence improves. Second, add to positions with credibility > 75 and size < 1% — under-sized high-evidence names are the inverse mistake; add if the position actually fits portfolio-level risk constraints. Third, portfolio-weighted credibility < 60 means the book is narrative-heavy — either the evidence improves soon or the book underperforms; no scoring system is a market timer but this threshold is a useful sanity check.

Where It Fits

Pair with the per-ticker Conviction Gauge, Conviction Components breakdown, and Portfolio AI Insights for natural-language drill-down into why specific positions are scoring low.

Open the Portfolio Credibility card → /app/portfolio