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 discusses why the differential equation for an American put may be written as an inequality in the continuation region above the optimal exercise boundary. One response links the inequality to the option’s delta at the boundary and the condition…
The document presents a replication attempt comparing price momentum with two earnings-based momentum signals, standardized unexpected earnings (SUE) and earnings announcement returns (CAR3). The researcher describes building SUE from quarterly earnings per…
The document describes using quasi-Monte Carlo (QMC) points to generate Gaussian shocks for simulated asset paths, with a Brownian bridge (BB) arranging those shocks across time. It recommends mapping low-discrepancy uniform points through the inverse normal…
The document explains why a historical adjusted closing price downloaded from Yahoo Finance can differ between downloads made at different times. The quoted example compares two values for Microsoft on the same historical date, then points to the provider’s…
The discussion explains how Fourier transforms and characteristic functions can price European options when the underlying log-price density is difficult to obtain directly. It starts from the discounted expected payoff, expresses that expectation as an…
The document raises the possibility that stocks listed early alphabetically might have different returns. The observation comes from a backtest comparing the first 10 and first 100 alphabetically ordered S&P 500 symbols; the first 10 appeared to perform…
The document explains how to price a zero-coupon bond at an intermediate time in the one-factor Hull–White model and debug a QuantLib simulation. The bond value is expressed as a time-dependent factor multiplied by an exponential function of the current…
The document explains winsorization as replacing observations in the upper or lower tails of a variable with the corresponding percentile cutoff. It contrasts truncation, which marks tail observations as missing, and frames both as possible treatments for…
The document describes fitting a credit-rating transition matrix as a convex blend of a base matrix and a stress matrix. A single scalar weight determines the blend, and the proposed objective is the sum of squared cell-by-cell differences from an observed…
The document asks how to apply the FDR+ method to select trading rules that appear to outperform. It says the author has computed p-values using a stationary bootstrap, following a cited study, but provides no worked calculation, code, or detailed definition…
The document addresses a mismatch between an analytic price and a Monte Carlo estimate for a floating lookback put under geometric Brownian motion. The question gives a volatility, maturity, interest rate, time step, and spot, and reports that the simulation…
The document explains two ways to simulate the return of a fixed-weight portfolio under a normal-return assumption. One method samples directly from a univariate normal distribution using the portfolio’s expected return and variance. The other draws asset…
This troubleshooting note concerns an error raised by the rugarch package while running a rolling GARCH forecast with ugarchroll. The reported setup estimates an sGARCH model with an ARMA mean specification, then forecasts one step ahead over a rolling…
The document explains how a t-statistic can test whether a trading strategy’s average return differs from a null value, commonly zero. A small statistic indicates insufficient evidence to reject the null of no excess return. Because return observations can…
The document asks how to annualize a Sharpe ratio calculated from monthly returns when those returns are serially correlated. It notes that the usual square-root-of-time scaling assumes uncorrelated returns and raises a specific question about estimating the…
This note shows how to price a path-dependent lookback call in a two-period binomial model. A European payoff depends only on the terminal asset price, so the two paths that reach the same middle terminal node can be combined and their probabilities added. A…
The discussion considers whether Markov chain Monte Carlo is useful for Bayesian inference in quantitative finance. It describes potential applications to estimating hidden-process parameters, including stochastic volatility models, and to inference for…
The document compares annualized return and volatility calculations based on logarithmic returns and simple returns. For monthly log returns, it describes scaling the mean by the number of periods in a year and scaling standard deviation by the square root…
The document explains how to simulate two correlated asset paths under geometric Brownian motion. It starts with a covariance matrix encoding each asset's volatility and their correlation, then factors that matrix with a Cholesky decomposition. Multiplying a…
The document presents a Python example that simulates portfolios by drawing random asset weights, scaling them to a fixed total, and calculating annualized return, volatility, and a return-to-risk ratio from historical returns and covariance. It asks how to…
The document addresses how to assess an exotic option pricing model when direct market prices are scarce. It proposes first checking whether the contract can be reduced to a simpler exotic structure that has observable quotes; if so, the model should…
The document asks whether stock indicators such as moving averages, MACD, Bollinger Bands, and exponential averages should use adjusted closing prices when shares split or pay dividends. Its example compares a two-day moving average around a reported…
The document examines a simulation of stock prices under geometric Brownian motion with zero drift. Although each simulated price at a fixed horizon has a lognormal distribution, the author finds that taking the median across all time steps produces a value…
The document gives formulas for calculating a daily pivot point and the first three support and resistance levels from the prior period's high, low, and close. It shows these levels alongside price bars and existing MACD and moving-average signals, then asks…