Quantitative Markets Research Lab — Findings https://kylewisniewski.com/lab/feed.xml 2026-08-08T00:00:00-06:00 Quantitative Markets Research Labhttps://kylewisniewski.com/lab Evidence-backed decision briefs on quantitative models, their practical consequences, and their limits. Can an “optimal” portfolio stay optimal? https://kylewisniewski.com/lab/finding-portfolio-estimation.html 2026-08-08T00:00:00-06:002026-08-08T00:00:00-06:00 Why a maximum-Sharpe portfolio’s estimated advantage degraded—and what risk-based methods did and did not fix. Does a 95% risk estimate fail only 5% of the time? https://kylewisniewski.com/lab/finding-var-backtest.html 2026-08-08T00:00:00-06:002026-08-08T00:00:00-06:00 A rolling historical VaR forecast produced too many breaches, and the exceptions clustered. Will bonds diversify stocks in the next stress? https://kylewisniewski.com/lab/finding-diversification-regimes.html 2026-08-08T00:00:00-06:002026-08-08T00:00:00-06:00 Stock–bond correlation changed sign in the studied sample, exposing the regime assumption inside diversification. Does a factor fund deliver the exposure on its label? https://kylewisniewski.com/lab/finding-factor-etfs.html 2026-08-08T00:00:00-06:002026-08-08T00:00:00-06:00 The advertised exposures appeared, but no alpha survived the multiple-testing threshold. Can one volatility number describe every option? https://kylewisniewski.com/lab/finding-volatility-smile.html 2026-08-08T00:00:00-06:002026-08-08T00:00:00-06:00 Observed volatility, fat tails, and an implied-volatility skew reject the constant-volatility Gaussian fingerprints. How many simulations are enough? https://kylewisniewski.com/lab/finding-monte-carlo-error.html 2026-08-08T00:00:00-06:002026-08-08T00:00:00-06:00 Monte Carlo accuracy depends on a declared tolerance, sampling variance, and the numerical bias the path count cannot cure.