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 considers turning an investor’s discretionary stock selection into a systematic screening process. Candidate inputs include institutional holdings disclosures, repurchases, insider trades, expectations, fallen-angel ratings, enterprise value…
The document discusses how to choose parameters for simulating stock prices with a Variance Gamma process. The question describes generating a gamma-distributed time change, then a conditionally normal increment, and asks how to estimate the process…
The document describes an out-of-sample comparison of one-day-ahead variance forecasts from an ARMA-GARCH model against realized-volatility estimates for twelve stocks. It proposes checking calibration by regressing observed variance on forecast variance: an…
DGTW adjusted returns measure a stock’s performance relative to a characteristic-matched benchmark. For each stock, the method subtracts the return of a portfolio of firms assigned to matching groups based on market equity, market-to-book ratio, and prior…
The document distinguishes the return between two prices from the total return generated by a sequence of long positions. It asks whether adding percentage returns from separate trades gives the correct overall result and whether a charted calculation method…
The document addresses event studies where many companies experience events on different calendar dates. Its central method is to align each event as time zero, estimate the normal-return model in a pre-event window, and express returns on surrounding days…
The document surveys ways to capture, store and later analyze market tick data. It describes collecting Interactive Brokers market updates through an R interface and writing records to files or in-memory time-series objects. For longer-term analysis,…
The question asks whether Markowitz portfolio construction should use closing prices or dividend-adjusted prices when estimating a minimum-variance portfolio, an optimal risky portfolio, and an efficient frontier. The answer frames the choice around how the…
The document discusses a coding question about simulating asset prices in a Least Squares Monte Carlo (LSM) method. The question focuses on formulas used to calculate the conditional log-price mean and volatility at an earlier time step, and asks why the…
The discussion addresses the intuition that reversing every long and short signal from a losing strategy should produce a winner. It explains why this does not follow: trading costs affect both directions, and a strategy’s rules for entering and exiting…
The discussion compares ways to obtain historical market depth data for research, including specialist vendors, broker APIs, trading platforms, and collecting a dataset over time. It distinguishes consolidated quotes from full-depth exchange feeds, since…
The discussion distinguishes quantitative trading from technical analysis by the rigor and breadth of the research process, rather than by how complicated an indicator is. A moving average or signal-processing filter can be used by either a discretionary…
The document explains why swaps manually priced from a QuantLib curve may fail to reproduce the market quotes used to build that curve. In the example, the discrepancies arise because the manually constructed swaps do not exactly match the swap rate helpers’…
The document considers portfolio rebalancing when expected-return signals refresh quarterly but risk estimates update daily. At an interim rebalance, prices have moved while the original alpha forecasts grow stale, creating uncertainty about whether to…
The document considers comparing two portfolio strategies when Monte Carlo simulation produces a collection of Sharpe ratio estimates for each strategy. The response uses an approximate normal sampling distribution for an estimated Sharpe ratio, with…
The discussion addresses why a Monte Carlo estimate of a Bermudan option value can fall below the true value when continuation values and exercise decisions are estimated from simulated paths. The central explanation is the supremum-estimation effect:…
The document considers how to normalize daily trading volume and value for stock time-series models, including recurrent neural networks. It contrasts whole-series statistics such as global minimum, maximum, mean, and standard deviation with normalization…
The document concerns locating the S&P 500 membership list for each day from 1960 onward, a data need relevant to historical index analysis. The answer suggests using a Bloomberg terminal and checking whether a university provides access to financial data…
The document examines why published annualized returns, standard deviations, and Sharpe ratios can be difficult to reproduce. It raises choices that affect the calculation, including sampling frequency, price timing, and simple versus logarithmic returns,…
Financial prices often behave as nonstationary series: their statistical properties can depend on the price level and past observations. Returns, calculated as percentage changes or differences in log prices, are more often treated as stationary. That makes…
The document presents a Python workflow for exploring portfolio risk and return from a table of asset returns. It generates random long-only weights that sum to one, estimates annualized portfolio return and volatility from sample means and covariances, and…
The document reports a historical data issue in Interactive Brokers requests: weekly and monthly bars for an IBIS-listed stock are labeled one day later than expected, while daily bars for the same contract and bars for NYSE appear correctly dated. The…
This note asks whether log differences are an appropriate transformation for a trend-stationary series modeled as a linear trend plus random noise. It contrasts examples where the series has a relatively large baseline and modest trend and noise with cases…
This note points readers to a study examining how stop-loss and profit-taking rules affect return distributions. It reports that stop-losses can reduce position volatility, but that hidden costs may offset the perceived benefit. The cited study finds no…