Research-Roundup
- This week in research: the reading is not the reality
New research on visibility-biased risk data, product labels covering different contracts, and why a good price fit is not a recovered risk distribution.
- This week in research: the gap between a prediction and a working rule
New research on retail signal families, machine-learning timing models, quantum kernels and accounting anomalies, and what an edge must survive to count.
- This week in research: the machinery underneath the number
New finance research on the plumbing behind measured results: momentum from cash mechanics, shifting bond correlations, AI backtest bias, and defaults.
- This week in research: risk lives in the relationships, not the labels
New research on measuring portfolio risk from the inside: risk as internal covariance, rotating diversification, counting real factors, and honest uncertainty.
- This week in research: why a measured edge is not a durable one
New finance research on why measured edges fade: trend-following's decline, factor-model gaps, volatility regimes, and evidence that prices still forecast fundamentals.
- This week in research: when the model is tidier than the market
New finance research on the limits of models: complex forecasts rarely beat simple ones, simulations miss market memory, and trading costs reward patience.
- This week in research: hundreds of factors, a handful of real bets
New finance research on what actually drives returns: hundreds of factors collapse to a few distinct forces, a global dividend premium, and the fat tails returns carry.
- This week in research: the false comfort of a strong track record
New finance research on why a strong track record flatters: Sharpe-ratio selection bias, unknown expected returns, unrealistic simulated paths, and the behavior gap.
- This week in research: why portfolio risk won't sit still
New research on how portfolio risk actually behaves: long memory, switching volatility regimes, correlation-driven diversification, and the measured behavior gap.