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 discussion considers how to select the orders of an EGARCH model for forecasting volatility associated with S&P 500 option portfolios. It distinguishes choosing a model for in-sample fit from choosing one for predictive performance. AIC, BIC, and…
The document asks whether a rolling Value at Risk backtest is appropriate for returns from a winner-minus-loser (WML) momentum strategy. The described procedure estimates a mean and variance from a moving historical window, computes a normal VaR threshold,…
The document asks how Cover’s universal portfolio algorithm compares with two simple allocations: a constant-relative-proportions portfolio that rebalances to equal weights, and a buy-and-hold allocation that initially divides wealth equally across assets.…
This exchange clarifies how to interpret labels in the triple barrier method. Each labeled instance corresponds to a specific trading circumstance and receives a single outcome according to which event closes it: an upper barrier win, a lower barrier loss,…
The document investigates why a return computed from a vendor’s adjusted closing prices differs from a hand-calculated dividend-inclusive return. The explanation focuses on the ex-dividend date: the investor’s daily return includes both the new share price…
The document considers how to simulate a variance process that can turn negative under a discretized stochastic volatility model with a variance-gamma jump term. Rather than forcing the jump component itself to be nonnegative, the answer presents ways to…
The document asks whether GARCH(1,1) parameters can be updated as each new price arrives, without rerunning the full estimation. Its answer is that maximum likelihood estimates depend on the full likelihood across observations. Because the GARCH recursion…
The document asks whether daily open, close, high, and low prices for thousands of US-listed stocks can be downloaded in one low-cost bulk file, without querying each symbol separately. The responses suggest using Python's Pandas DataReader to retrieve…
The document considers how to calculate investment performance when capital is added to or withdrawn from an account during the measurement period. A simple comparison of ending balance with starting balance can misstate strategy performance because it…
The document explains two ways to calculate an annual return from daily prices: compute the simple return over a date interval, or compound each daily return by multiplying one plus each return and subtracting one. It also mentions log price differences as a…
The document explains how to simulate a down-and-in barrier call under a local volatility model. At each simulation step, the volatility should be evaluated at the current time and stock level, rather than at the option’s fixed strike and maturity. The…
The note distinguishes the roles of the ARMA and GARCH components in a return model used for directional trading. The ARMA conditional mean supplies the expected return and therefore drives a rule that goes long or short according to the forecast sign. GARCH…
The document raises a backtesting problem for a systematic strategy that repeatedly buys one-year, 90% strike equity index puts and rolls them shortly before expiry. It proposes repricing the option each month with Black–Scholes using the current spot,…
The document compares historical daily log-return standard deviations for NASDAQ, gold, and silver with average standard deviations measured across simulated GARCH(1,1) paths. The simulations use rugarch, fit separate models with a zero-mean specification,…
The document considers how to assess expected shortfall, also called conditional value at risk, when it is not directly elicitable as a standalone forecast. It discusses a proposal to approximate expected shortfall by averaging several value-at-risk…
The document examines whether a cross sectional mean reversion strategy can replicate an equity index using a subset of its components. The proposed signal buys stocks whose returns lag the universe and shorts those that outperform. The response explains…
The document considers Monte Carlo evaluation of conditional expectations involving an integrated Jacobi volatility process, as part of a real-options research problem. The process is bounded, but the questioner worries that Euler discretization may…
The document asks how to obtain a discrete-time form of the Jamshidian–Zhu short-rate model, which combines an Ornstein–Uhlenbeck stochastic driver with principal-component weights for yield-curve movements. The stated application is estimating stressed…
The document presents a simple moving-average backtester for shares, with signals based on a fast and slow average, position sizing, cash accounting, and portfolio equity. The reported output contains missing values, and the answer points to a likely cause…
The document asks how to adjust statistical significance when a trading rule is selected by optimizing continuous parameters, using moving-average lengths as an example. It contrasts this with calendar-effect research, where a finite set of candidate rules…
The document explains why the estimated constant in a GARCH return model may differ from the ordinary sample mean. The sample mean gives every observation equal weight. Under a GARCH model, observations have different conditional volatility, so…
The document outlines a practical workflow for stock factor investing: define an investible universe, collect and process factor data, then rank or score securities. It describes several ways to turn scores into holdings, including selecting a subset and…
The document compares Monte Carlo estimates for discretely monitored variance and volatility swaps under a Variance-Gamma model. It defines the variance swap value as the square root of expected realized variance, while the volatility swap value is the…
The document discusses how to evaluate a market forecasting model when each observation at time T is paired with an outcome measured n days later. Randomly assigning calendar years to training and testing sets can create boundary problems because future…