Markowitz Mean-Variance

Markowitz mean-variance optimization treats portfolio construction as a quadratic program: given a vector of expected returns and a covariance matrix, choose weights that maximize expected return for a given variance, or minimize variance for a target return, under constraints like full investment and no shorting. The machinery is elegant and the frontier it traces is genuinely efficient, but only relative to the inputs you fed in. Because the optimizer exploits differences between assets aggressively, it amplifies estimation error, tilting toward assets whose returns were overestimated or whose risk was underestimated. That instability is the founding problem of this track, and every later node is a remedy for it.