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 how to weight a small set of stocks to track the S&P 500. One approach is to estimate weights from historical returns, seeking strong correlation with the index or low tracking error. Regression, principal component methods, and…
The document asks how to backtest and plot a strategy whose signals indicate long, flat, or short exposure, using historical prices. The proposed calculation forms log price returns, multiplies them by the prior period’s signal to represent the position held…
The document raises practical questions for researching and backtesting crypto or forex strategies when exchange pairs use different quote currencies. It asks how to value holdings in a common currency, such as USD, whether to use a direct market or a chain…
The document gives Bloomberg-based routes for locating swap rates and historical swap data, with GBP as its example. It recommends using the Curve Finder to filter for GBP, entering a tenor-specific swap query such as a five-year GBP swap and following the…
The document responds to a request for news archives extending beyond a short recent window. It names three possible sources: AYLIEN’s historical news API, Bloomberg news and sentiment data used in historical backtesting, and Webhose.io, which an answer says…
The document considers whether signals from a long-only equity momentum system could be used to sell puts. One response suggests choosing a high-momentum stock using a risk-adjusted return measure and selling a near-the-money put with a later expiration,…
The discussion clarifies how to choose a time increment when simulating stock prices with geometric Brownian motion. For a daily model expressed in annual units, the drift scales with elapsed time and volatility with the square root of elapsed time. A step…
The document asks whether leverage should multiply an index or ETF’s tracking difference by the leverage factor. The response explains that this equality need not hold when leverage is applied each day or over another subperiod, because compounded returns…
The document examines what an annualized Sharpe ratio can imply about the chance of a losing year. It standardizes annual PnL by its mean and standard deviation, then uses a symmetry assumption and Chebyshev’s inequality to derive an upper bound. For the…
The document describes an effort to reproduce parameter estimates from continuous-time stochastic models of US inflation. It lists reported estimates for log-normal models with constant, local, and mean-reverting Vasicek drift, using historical CPI data and…
The note surveys ways to approximate a hedge fund’s return stream using a different, potentially passive strategy. It groups prior research into linear factor models, nonlinear or less transparent models, and work emphasizing the difficulty of reliable…
The document considers how to fit a predicted distribution of future stock log returns using the realized profit and loss of European calls and puts at expiration. It asks how this objective translates mathematically and compares it with likelihood-based…
The document asks how to evaluate exponential moving averages as a forex trading strategy and requests research on their effectiveness. The response offers two starting references on technical analysis in foreign exchange: a 1992 study of technical analysis…
The document discusses the edge ratio as a tool for evaluating the entry component of an algorithmic trading system. It presents two limitations: unlike the Sharpe ratio, the edge ratio is not commonly requested by investors, and its statistical properties…
The document compares estimating portfolio value at risk from a multivariate normal model with using historical returns or a copula. Historical simulation applies the observed joint return series directly, so it can reflect cross-asset dependence and…
The post considers how to allocate capital across more than fifty backtested strategies that trade ETFs, mostly equities, with daily signals and next-day open execution. The strategies are mainly short-horizon mean-reversion systems, with different trading…
The document asks whether a proposed method is suitable for evaluating an equity trading strategy when only about ten years of historical returns are available. The researcher resamples the returns with a stationary bootstrap, evaluates candidate strategies…
The document explains that adjusted historical prices incorporate corporate actions that occur after each observation date. Because later dividends and other events are applied cumulatively, older adjusted closes can differ increasingly from the raw closes…
The document discusses whether Interactive Brokers can provide enough historical intraday data for statistical modeling, using S&P 500 exposure as the example. It points to the R IBrokers package and its historical data request function, which can specify an…
The document asks how to measure monthly returns for two S&P 500 equity portfolios formed from the highest- and lowest-scoring ESG companies. The proposed portfolios are reconstituted annually using the prior year’s ESG scores, and the author wants to…
The document describes estimating incremental value at risk by calculating portfolio VaR from simulated scenario returns, removing one position’s scenario returns, and recalculating VaR. Incremental VaR is the difference between the original and reduced…
The document examines why win rates become ambiguous when a strategy accumulates or liquidates positions over multiple transactions. It gives a proposed comparison based on total capital invested, total proceeds, and the elapsed time from the first purchase…
The document compares a Python Least Squares Monte Carlo (LSM) implementation for American options with QuantLib pricing engines. The answer emphasizes that LSM continuation-value regression should use in-the-money paths, and that regression is an…
The document addresses why repeated geometric Brownian motion simulations produce different price paths, even when the model inputs remain unchanged. The randomness comes from newly drawn Brownian shocks on each run. The accepted response gives a practical…