{
 "spec": "wiki-changes/1",
 "site": "quants.wiki",
 "origin": "https://quants.wiki",
 "reviewed": "2026-08-27",
 "generated": "2026-09-03",
 "note": "Dated content-change events, newest first. Poll this instead of re-crawling.",
 "entries": [
  {
   "date": "2026-08-27",
   "type": "created",
   "url": "https://quants.wiki/",
   "note": "Initial publication: 76 estimator entries and 25 reference tables across six sections, built on three published datasets so the whole corpus is independently checkable - a 24-month return series with a benchmark and a constant risk-free rate, ten daily OHLC bars, and a three-asset covariance matrix. Performance statistics: the Sharpe ratio with the Lo 2002 autocorrelation correction, the iid and non-normal standard errors, the Miller and Gehr small-sample bias factor, Sortino under both downside-deviation divisors, Calmar, three incompatible Sterling definitions, Omega, the information ratio against the appraisal ratio, Treynor, Jensen's alpha with its t-statistic, M-squared, up and down capture under both conventions, and the three annualisation rules that disagree. Drawdown: maximum drawdown, average drawdown, Ulcer index, episode duration and recovery under censoring, time under water, the expected maximum drawdown of a driftless random walk, and why maximum drawdown is a biased risk estimator. Estimation: Marchenko-Pastur eigenvalue bounds by N/T, the condition number and what it costs, the positive-semi-definiteness constraint on pairwise correlations, Ledoit-Wolf shrinkage and the closed form for a target condition number, EWMA and GARCH(1,1) with their parameter constraints, and the close-to-close, Parkinson, Garman-Klass, Rogers-Satchell and Yang-Zhang estimators computed on identical bars. Portfolio: mean-variance, tangency and minimum-variance closed forms, two-asset closed forms with a sensitivity table, the Euler risk-contribution identity, equal risk contribution against inverse volatility, hierarchical risk parity, Black-Litterman, reverse optimisation, Kelly and fractional Kelly, and the transaction-cost no-trade band. Backtest validity: the expected maximum Sharpe ratio from N trials, the probabilistic and deflated Sharpe ratios, minimum backtest length, how many trials produce a spurious Sharpe of 2, the probability of backtest overfitting, look-ahead and survivorship bias, overlapping labels and the effective sample size, purged k-fold cross-validation, the embargo, and walk-forward against combinatorial purged cross-validation. Execution: implementation shortfall decomposed into delay, impact, opportunity cost and fees, the effective, realised and quoted spread identity, the square-root impact law and the capacity it implies, the Almgren-Chriss trajectory, VWAP tracking error, and Kyle's lambda. Added performance statistics and backtest deflation calculator descriptors."
  },
  {
   "date": "2026-08-27",
   "type": "corrected",
   "url": "https://quants.wiki/performance/",
   "note": "Non-normal standard error of the Sharpe ratio corrected from (gamma4 - 3)/4 to (gamma4 - 1)/4, the canonical Mertens numerator. The file contradicted itself: its probabilistic-Sharpe entry already used the correct form. Monthly SE moves from 0.216692 to 0.220230 and annualised from 0.750643 to 0.762899."
  }
 ]
}