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 explains how to test whether a signal predicts returns by sorting securities into portfolios according to the signal, then comparing high-signal and low-signal groups. A spread portfolio is formed by going long the high group and short the low…
The document considers which market variables might be studied alongside Twitter activity, including index and stock prices, price differences, trends, and expected returns. A response recommends treating posts as a source of market sentiment rather than…
This exchange considers how to update expected asset returns when rebalancing a Markowitz portfolio day by day. The question proposes incorporating each new observation into the historical sample and recalculating the expected returns and portfolio weights.…
This document explains how to calculate the probability that a geometric Brownian motion (GBM) price will fall inside a specified interval at the next time step, given its current value. This conditional probability is used in the stochastic mesh method to…
The document asks why Jegadeesh and Titman used overlapping portfolio holding periods when testing momentum strategies. The replies give two intuitions: overlapping periods can create a larger sample than non-overlapping periods, and examining portfolios…
The document describes an attempt to estimate Ornstein–Uhlenbeck process parameters by maximizing a Gaussian transition likelihood, then asks why the estimates fail or produce a math domain error. The replies identify a numerical singularity: the conditional…
The document explains how to interpret a cross-sectional rank correlation used in a trading alpha. At each date, rank securities by each variable across the same universe, such as open price and volume, then calculate the correlation between those two…
The document advises aspiring practitioners to learn statistics and machine learning before applying models to market data, then study financial modeling and backtesting, including transaction costs. It explains model fitting in terms of prediction error and…
The document explores how to backtest an Avellaneda–Stoikov style market-making strategy using historical limit order book data. The question describes estimating order-arrival intensities from synchronized order and book events, then asks how simulated…
The document raises two questions about using combinatorial purged cross-validation (CPCV) for a binary trading model: whether to select hyperparameters by averaging metrics across splits or across reconstructed paths, and whether evaluating those parameters…
The document asks how to compare a numerical European call price from an explicit finite-difference scheme with the analytical Black–Scholes value at an intermediate time step. The model uses spatial grid points corresponding to underlying asset prices, and…
The document addresses why fitted AR parameters from a simulated ARMA series may differ substantially from the values used to generate it. The response says the simulation and fitting logic are sound and points to sampling variability: a series with only a…
The document considers whether publishing a technical indicator can reduce its subsequent trading performance. It frames a strategy’s historical results as a mix of genuine alpha and overfitting, with only genuine alpha expected to persist out of sample;…
This discussion derives the coefficient that minimizes variance when estimating an expectation with a control variate. For an adjusted estimator of the form X plus θ times the difference between the known mean of Y and Y, the constant mean contributes no…
The document shows how to plot fitted conditional volatility together with forecasts from a GARCH(1,1) model in R. Its example uses the fGarch package to specify and simulate a normally distributed GARCH process, fit the model, and predict future conditional…
The document discusses discretizing a Merton-style jump-diffusion process for simulation. Rather than applying a basic Euler step directly to the asset price, the accepted answer works with log price. Each interval adds the drift adjusted for diffusion, a…
The document introduces the information available in a visible order book: current prices, displayed quantities, and the structure of supply and demand. Its question asks how that information could support arbitrage or manipulation, but the accepted answer…
The document explains how to build a simplified exchange execution simulator for backtesting. It recommends using historical order book snapshots or updates to estimate fills: a market order consumes available depth until its requested quantity is filled,…
The document asks how to compute the hindsight-optimal sequence of holdings across a universe of stocks when an investor can trade daily, compounds returns, pays transaction fees, and cannot short. It considers dynamic programming and presents an attempted…
The document describes a strategy using 30-second bars on E-mini S&P 500 futures that performed well in backtests and walk-forward tests but lost in live simulation because small execution-price differences erased its edge. The question asks whether…
The document discusses how to reproduce Excel's bond YIELD calculation with QuantLib. QuantLib's bond yield method takes a target price, day-count convention, compounding convention, and coupon frequency, while settlement, maturity, coupon schedule,…
The document explains the practical limits of timestamp precision in stock market data, distinguishing timestamp resolution from the frequency of actual trades and quote updates. Nanosecond timestamps are used by specialized data providers and can help order…
The document examines an apparent mean-reverting basket built from several foreign-exchange rates. The proposed process fits a linear regression to estimate hedge ratios, forms a spread, then applies an augmented Dickey–Fuller test. The reported significance…
The document poses a valuation problem for a forward-starting interest-rate swap in a binomial short-rate lattice. The swap begins after the valuation date, receives floating payments, pays a fixed rate, and makes payments in arrears. The questioner reports…