Statistics
The statistical results a backtest relies on, each with its formula, its source and what it changes in practice.
What this shelf covers
The statistics a strategy evaluation relies on, each stated with its formula and its assumptions: the standard error of a Sharpe ratio under autocorrelation, bootstrap schemes for dependent returns, tests that compare two forecasts or two Sharpe ratios, shrinkage of noisy estimates, and first-passage probabilities for drawdown limits. The worked numbers can be recomputed by hand or with the calculators on this site.
12 notes
Regime Changes in Financial Markets (Ang-Timmermann Review)
Why Prices Are a Martingale: Bachelier (1900) and Samuelson (1965)
Correlation Matrices When Assets Outnumber Observations
Efron's Bootstrap: Resampling the Sample to Get a Standard Error
Extreme Value Theory for Tail Risk
Volatility Models: GARCH, Realized Volatility, HAR
Granger Causality Measures Prediction, Which Is Weaker Than Cause (Granger, Pearl)
Pesaran-Timmermann Test of Directional Accuracy
The Standard Error of a Sharpe Ratio, and What Autocorrelation Does to It
Slutsky Effect: A Moving Average of Noise Manufactures Cycles