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 using proxy time series when a risk factor lacks enough historical observations for market risk measurement. It contrasts older Basel II guidance with the newer Fundamental Review of the Trading Book (FRTB), where insufficient real…
The document describes a two-period portfolio comparison using historical returns. It estimates an in-sample efficient frontier from the first period’s empirical covariance matrix, generating portfolios across target returns. It then keeps those portfolio…
The document explains why historical OHLCV records for the same listed share can disagree across Google Finance and Yahoo Finance. Possible causes include distinct vendors and trading platforms, different aggregation rules, post-trade corrections,…
The document asks whether stock price paths can be simulated with different models while holding the initial price and historical mean and standard deviation fixed. It names geometric Brownian motion (GBM), Merton jump diffusion, and Heston stochastic…
This discussion surveys Python approaches to portfolio construction, focusing on mean-variance and minimum-variance optimization. It points to convex optimization packages that can solve quadratic portfolio problems, with higher-level interfaces for…
The document describes a one-expiry SPX implied-volatility calibration using the SSVI parameterization. The initial implementation produced implausible estimates when fitting market option prices, and the author also found that the fitting objective had…
The document poses a research question about whether the choice of volatility model can predict Value at Risk backtest outcomes in advance. It asks whether a GARCH-based VaR forecast, for example, should be more likely than an ARCH or unconditional forecast…
The discussion asks how to construct a discrete-time counterpart to a continuous diffusion whose state remains between zero and one. A direct update using normal noise may leave the interval because Gaussian increments are unbounded. One response proposes a…
The document addresses how to test whether two indicators help explain a time series of returns when the indicators may be autocorrelated or nonstationary and a VAR analysis has not produced significant results. Its proposed starting point is a linear…
The document compares White’s Reality Check with the Benjamini-Hochberg-Yekutieli procedure for screening a large set of backtested strategies against a benchmark. It describes the Reality Check as accounting for dependence among test statistics through…
The document describes diagnosing weak convergence in Monte Carlo simulations of an option-pricing model with state-dependent volatility. The original comparison estimates expectations at two time-step sizes, but uses independent simulated paths. Sampling…
The document examines claims that artificial neural networks outperform Black–Scholes in option pricing. The cited study generates Black–Scholes estimates using simplified inputs, including a closing spot price, fixed volatility and interest rate, no…
The document explains why a backtest based only on the best bid and best ask can overstate achievable execution quality. Those quotes represent the most favorable available prices for limited quantities; larger orders may consume that liquidity and fill at…
The document explains how to represent a Wiener process on a uniform time grid for simulation. Divide the interval into equal steps, then construct each sampled value by adding the next increment to the previous value, starting from zero. The increment over…
The discussion considers how to allocate capital to an automated crypto strategy whose backtest suggests substantial drawdowns, and when to scale its exposure or stop trading as its edge may fade. The response emphasizes that changing position size is a…
The document asks whether a time-series model that chooses a fully invested long or short position could be trained by maximizing simulated terminal portfolio value instead of minimizing squared prediction error. It proposes simulating daily positions from…
The document considers how to build a weighted index from several equity index series when their market calendars do not match. It identifies two basic approaches: restrict calculations to dates common to all series, or carry forward the previous available…
The document asks whether bootstrap resampling should happen before strategy data mining or after a candidate strategy has been selected. It describes a monthly SPY and TLT rebalancing strategy tested over six years and 72 trades, reporting a backtest Sharpe…
The document discusses a lower-cost approach to storing and querying market tick and bar data when specialized financial databases are too expensive. The response suggests using an OLAP server with cube-based aggregation. It describes persisting ticks at…
The answer recommends separating operational monitoring from strategy monitoring when running automated trading systems. Operational checks include server heartbeats and alerts, centralized log persistence, and visibility into CPU and memory use. It…
The document addresses where to obtain historical European Union Allowance futures data at intraday frequency for carbon market research. The response points to a commercial market data vendor that offers trade histories and, optionally, quote histories for…
The document explains why a historical portfolio Value at Risk calculation can return an empty result at a high confidence level. The cited implementation converts the confidence level to a lower-tail probability, multiplies that probability by the number of…
The document describes an LSTM trading model using RSI, MACD, a ratio of two exponential moving averages, and recent returns as inputs over a short history. It assigns separate long and short outputs based on whether a hypothetical trade reaches a…
The document describes a practical signal-processing problem in a trading system that receives live bid and ask data. The author reports that offline backtests can use non-causal smoothing methods with access to future observations, but those methods lag…