The Black-Litterman model solves mean-variance optimization's biggest problem: extreme sensitivity to expected return inputs. Small changes in return assumptions cause wild weight swings. Black-Litterman starts with market-implied returns (reverse-engineering the CAPM equilibrium) as a neutral baseline, then tilts toward your views (e.g., 'I think EM will outperform by 2%'). Key innovation: Uses Bayesian updating to blend market equilibrium with your views, weighted by confidence. This produces stable, diversified portfolios instead of concentrated bets. Output: Expected returns that balance market consensus and your insights. Adoption: Widely used by institutional investors. Limitation: Still requires subjective view inputs, but handles them more gracefully than raw mean-variance.