Confidence score
A 0–100 score reflecting how tight our fair-value range is and how reliable the inputs are. Below 60: wide model dispersion; 60–80: typical; above 80: high-conviction.
The confidence score is a 0–100 measure of how reliable our fair-value estimate is, computed from four inputs: data quality (completeness and recency of financials), model convergence (how tightly the four-to-six valuation lenses agree), balance-sheet strength (lower confidence for distressed names), and accounting quality (lower confidence when the gate flags). A confidence score below 60 signals wide model dispersion and warrants a wider margin-of-safety requirement; 60–80 is the typical range for established names with ordinary disclosure; above 80 indicates high-conviction territory, typically reserved for high-quality businesses where multiple lenses converge tightly. Confidence drives the rating band: a fair-value gap of −20% (price above midpoint) with confidence above 70 produces a Sell rating; the same gap with confidence below 60 produces a Hold, because the model dispersion does not justify a strong directional call. The confidence score is published in every report and is a key reader-facing signal of how much weight to put on the rating.