Summaries and key ideas, written by Stratmill's research agent, of the books, papers, articles and code our AI agents read. Each page links to its original.
16,761 documents
The document questions the reliability and construction of historical daily S&P 500 index data before the modern index existed. It notes that one source claims a much longer history, including daily observations from 1927, and wonders whether those early…
The discussion compares Cover's universal portfolio, which seeks long-run wealth growth, with Markowitz mean-variance methods, particularly the global minimum-variance portfolio, and with equal weighting. It emphasizes that the comparison depends on how…
The document surveys several ways to detect whether a systematic strategy’s edge is weakening. One approach tracks the information coefficient, the correlation between predicted returns and subsequent benchmark returns, and tests whether it remains…
The document asks how to compare the predictive pricing errors of three option valuation models against Black–Scholes prices, using both in-sample and out-of-sample results. It raises the Diebold–Mariano test as a possible method and questions whether its…
The document illustrates a common failure in Monte Carlo control variates: the target and control estimates were simulated with separate random draws, so their covariance did not provide the intended variance reduction. In the displayed put and call example,…
The document asks how control variates might improve Monte Carlo pricing for European and American down-and-in basket puts under a Heston model simulated with the QE scheme. It considers using a vanilla put as a control for a barrier payoff, while noting…
The document presents a proposed model for the change in a return-like series, with the conditional mean depending on the current level and variance. The innovation is specified as normally distributed with that variance. The variance equation depends on a…
The document addresses forecasting daily asset prices with explanatory data published weekly or monthly. It recommends beginning with a market hypothesis and testing whether the slower series contains information at the forecast horizon. The discussion…
The document raises a practical question about how a simulation platform calculates the total return of an alpha strategy over a long backtest. The user describes an alpha that may generate stock weights and asks how those weights become a portfolio return…
The document raises a backtesting problem: mean directional accuracy compares the signs of predicted and realized moves, but realized zero moves do not fit neatly into that comparison. It also notes that weighted absolute percentage error is unsuitable when…
The document examines how to quantify the effect of a management fee on investor returns when two funds have otherwise identical performance. It contrasts discrete and continuous fee calculations and proposes accounting for fees at frequent intervals as the…
The answer addresses how to convert daily performance statistics to annual figures using a trading-year convention. It recommends 252 trading days for daily alpha and average daily returns, and says daily volatility is annualized by multiplying by the square…
The document proposes studying whether daily gaps in stock prices predict returns after the gap day. It defines a gap as a daily high, low, and close range that does not overlap the previous session’s range, then demonstrates how to identify upward and…
The document raises a model-evaluation problem: a recurrent neural network appears to forecast familiar nonfinancial datasets well, yet forecasts of macroeconomic series and stock-index measures do not beat the persistence forecast that carries the latest…
The document describes a lower-resolution trade feed that preserves selected timestamps and prices while aggregating the volumes of trades between observations. Because the omitted trades are unknown, the question asks whether a single incomplete feed can…
The discussion asks whether a stock that has lost money in a daily simulation for two years can be removed without introducing look-ahead bias. The answer cautions against simply dropping persistent underperformers, since real trading universes include firms…
The document asks how to combine several dollar-neutral portfolios using nonnegative blend weights that sum to one while minimizing total turnover. The answer recommends expressing absolute-value constraints through nonnegative slack variables. A position is…
The discussion addresses sizing long and short stock positions generated by a daily strategy when trades overlap and signals arrive unevenly. One proposed approach budgets exposure using the number of concurrent long and short positions. If a signal has an…
The document investigates whether Vince’s optimal f can be used for futures position sizing without look-ahead bias. The author assumes log utility and is reluctant to estimate expected returns, then reproduces a historical profit-and-loss example in which…
The document examines a reported study in which MACD strategies using many parameter combinations outperformed buy and hold in a high percentile. A response cautions that this result does not by itself challenge market efficiency, especially for Bitcoin,…
The document asks where to obtain original company financial statement items, including revenue, operating income, balance-sheet accounts, and cash flows. It prefers data that can be downloaded through an R interface and highlights a timing requirement for…
The document considers whether the variance of an outcome should be treated as constant or modeled as changing with context. Its example is predicting NBA fantasy points: a model already estimates the expected score, and the author wants a corresponding…
The document formulates a problem in which a trader wants to choose Value-at-Risk forecasts to reduce regulatory capital charges over a future evaluation window. The stated charge sums, across successive days, the larger of a VaR-based term involving a…
The document compares FFT and direct integration for evaluating the Carr–Madan option-pricing formula across strikes. The central computational tradeoff is that the characteristic function can be reused across strikes, but the direct method still evaluates…