What problem does it solve? Comparable-company analysis often reduces to picking a multiple and calling a stock cheap, ignoring claimholder mismatches, skewed distributions, and fundamental differences between firms. This Skill enforces Damodaran's four-test framework (definitional, descriptive, analytical, application) so every relative valuation verdict is controlled for growth, risk, and reinvestment rather than asserted. ## Core Features & Use Cases - Consistency-checked multiples: Computes PE, PBV, EV/EBITDA, EV/Sales, EV/IC and more, refusing mismatched pairings like EV/Net Income and negative denominators instead of returning meaningless numbers. - Distribution and regression analysis: Generates peer medians, quartiles and drop-out counts, locates a multiple in bundled market distributions, and fits OLS regressions in pure Python with t-statistics, R-squared, and multicollinearity warnings. - Sum-of-the-parts and cross-holdings: Prices conglomerate divisions against their own sector multiples, capitalizes unallocated corporate overhead, and values minority stakes and minority interests at market rather than book. - Use Case: Ask whether a stock trading at 8.9x earnings is cheap. The engine fits the sector regression of PE on growth and emerging-market risk, predicts 8.35, and reports the stock as 6.6% expensive despite its low headline multiple. ## Quick Start Ask the agent to price a company against its peers by running the multiples engine with the firm's market cap, debt, cash, and EBITDA, then control for fundamentals with a sector regression.
