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.
5,018 documents
The document explores estimating jump probability and jump volatility from an observed time series for use in a jump diffusion model. The proposed approach compares short- and long-window rolling standard deviations: a large divergence, especially near a…
The document investigates what people mean by the Sugihara Trading System. The response cautions that the term does not appear to denote a specific, practical trading system or a defined forecasting model. Instead, it points toward broader research on…
This exchange considers whether copulas can model dependence between variables with little or no linear correlation, and why a fitted conditional distribution may appear uninformative. The answer emphasizes that zero or near-zero correlation does not…
The document outlines a proposed workflow for estimating multivariate covariance under changing market regimes. It begins by fitting a GARCH(1,1) model to each series, removing estimated volatility, and fitting a hidden Markov model to classify observations…
This note considers whether a participant’s algorithm can be reconstructed from observed competition trades. The responses emphasize that trade streams alone leave many possible explanations because strategy decisions depend on hidden variables, market…
The document summarizes machine learning applications in quantitative finance, focusing on a study that uses recurrent neural networks with LSTM units to predict up or down price moves from high-frequency stock data. It reports that a model trained across a…
The document asks how to estimate parameters in models such as GARCH, Heston, and stochastic volatility with jumps, and whether standard methods like maximum likelihood or least squares can be applied. It also asks about general-purpose estimation without…
The document distinguishes a joint hypothesis about regression coefficients from tests of individual coefficients. The null that all slope coefficients are zero is typically assessed with an overall F-test; individual p-values address separate coefficient…
The document explores how to adapt a deep reinforcement learning portfolio model, originally formulated for long-only positions, to allow positive and negative weights. The proposed representation makes positive weights long exposures and negative weights…
The document describes a stock-return classification problem with a one-month horizon. The proposed labels distinguish negative and positive cumulative returns, with a threshold separating smaller from larger moves. Candidate predictors include lagged…
The document asks how a paper’s description of maximizing a likelihood by iterating between the Marquardt, or Levenberg–Marquardt, algorithm and BHHH should be understood. The practical distinction is whether each optimizer is run to completion before…
The discussion considers what additional time series might help a neural network forecast daily foreign exchange rates. Suggested inputs include currency crosses, interest rate differentials, yield curve slopes, and changes in rates, along with country-level…
The document asks how machine-learning model validation fits within walk-forward analysis or optimization. It contrasts choosing strategy parameters on a training period and measuring them on a test period with the machine-learning practice of reserving a…
The discussion addresses forecasting private equity fund capital calls and distributions, including whether forecasts can be related to market and economic conditions. It points to a theoretical reference on modeling illiquid alternative asset funds and…
The document describes an intended workflow for forecasting a crypto spread. The author first runs an OLS regression on first differences of time series, then treats the resulting residual spread as a tradable series. They inspect its autocorrelation and…
The document discusses simulating asset returns with non-normal marginal distributions while preserving dependence between assets. It notes that standard univariate samplers can generate Weibull or lognormal draws, and describes inverse transformation as a…
The discussion proposes treating corporate debt instrument choice and financing volume as related but separate modeling tasks. A researcher could first estimate the probability of issuing each debt type, then use issuance type as an input to a volume model,…
The discussion asks how to simulate observations from a state space model whose latent state follows an AR(1) process. It focuses on whether to discard initial observations as burn-in so the simulated process approaches its stationary distribution, and…
The document explores how to model volatility that rises sharply and then decays more gradually, producing an asymmetric sawtooth pattern. It starts from a stochastic volatility process with centered, standardized lognormal shocks, then considers a…
The document asks whether a power spectral density threshold can be selected dynamically when using the fast Fourier transform to denoise a signal. It compares this desired selection with retaining enough singular value decomposition components to capture a…
The document considers how to assess robustness when a trading strategy has several tunable parameters. It gives moving-average crossover lengths as an example and describes plotting backtest returns across two parameter dimensions as a heatmap. Broad areas…
The discussion considers whether robust intraday strategies can exist despite noisy price movements. It explains that a random-walk simulation is not a reliable way to discover technical trading signals, while risk controls such as profit targets and stop…
The document asks how to construct intraday returns whose aggregated daily returns follow a Student’s t distribution, contrasting this with the aggregation property of normal returns. One response points to Lévy processes as a framework that can include…
The document discusses how to choose a training duration for a deep reinforcement learning portfolio agent when training loss keeps falling but out-of-sample performance has flattened. It frames the number of training epochs as a hyperparameter and…