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.
8,851 documents
The document implements a Turtle-style breakout strategy using Donchian channel levels for entries and exits, with ATR-based stop placement and position sizing. It tracks the trade’s high and low, moves stops as prices advance, and adds to positions at…
The document raises the hypothesis that trend-following strategies may perform better in low-volatility conditions and asks how a trader might switch between trend and range-bound approaches. It defines volatility as the standard deviation of the price rate…
This post asks whether a proposed Average Directional Index calculator produces plausible values across several example price sequences. The examples include steadily rising and falling prices, an accelerating rise, a constant series, and a repeating…
The example explains why restoring target weights in a leveraged long-short portfolio can require buying an appreciated asset and increasing a short position in an underperformer. After a strong gain in the long asset, the portfolio’s total value has also…
The document asks whether volume-based bars can make fixed-parameter strategies more responsive to changing market activity. It starts from the observation that equal calendar periods may represent very different conditions, then suggests volume bars as a…
The document asks how to calculate a basic win percentage for a trend-following model that produces long and short signals. The proposed approach is to compare the price when a signal begins with the price when the signal changes, count a trade as a win if…
The document offers a reading guide for someone moving from foundational portfolio theory toward practical portfolio management. It starts with Markowitz’s portfolio selection work and Sharpe’s capital asset pricing model, then points to research on Bayesian…
The document considers when an investor should exit an underperforming hedge fund or private equity investment. It highlights practical frictions: exit windows can be long, fees may be substantial, and re-entry can be costly. It also notes the possibility…
The discussion diagnoses an implausibly large cumulative return in a simple moving-average crossover backtest. The strategy uses closing prices and a 10-day simple moving average to set a long or short position, then applies that position to daily asset…
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…
The document addresses how to identify and plot peaks in a stock price series. It notes that peak definitions vary and that a threshold parameter in a peak-finding function may be unclear. As an alternative, it describes a streak-based method: classify…
The document addresses how traders might identify market conditions associated with changes in algorithmic strategy performance. It proposes structural-break analysis as one avenue, naming CUSUM tests on recursive residuals or levels and explosiveness tests…
The discussion considers how Jim Simons may have recognized that market prices contain exploitable structure. Answers suggest that mathematical skill, curiosity about price patterns, and experience with signal detection can help researchers distinguish weak…
The document proposes an aggressive ETF rotation strategy inspired by dual momentum. Instead of comparing a small set of broad holdings, it would compare ETFs across asset classes, regions, sectors, and leveraged products, then select the strongest performer…
The discussion asks whether an asset is equally likely to trend or range and whether that likelihood can be calculated. The answer distinguishes trend from volatility and suggests that expected return, implied volatility, and the Sharpe ratio can help…
A hidden semi-Markov model allows the probability of leaving a hidden state to depend on time spent in that state, unlike a standard hidden Markov model’s memoryless transition structure. The document asks whether a trading model could also use cumulative…
The document considers whether a trend-following trader should express a signal on a base currency, such as GBP, through several currency pairs rather than through a single pair. The response argues that a basket of terms currencies may better represent…
The document proposes a repeatable way to identify liquid futures markets for research or systematic trend following. Rather than starting from a narrow published index or trying to query every contract through a terminal, it suggests parsing exchange…
The note considers why volatility can show long memory and why its measured persistence may vary across periods. It describes an analysis of daily realized volatility for multiple assets, comparing autocorrelations before and after the global financial…
The note raises two implementation questions about the time-series momentum method described by Moskowitz, Ooi, and Pedersen. First, it asks how to interpret the paper’s ex-ante volatility estimate when the strategy uses a twelve-month return signal: whether…
The discussion examines whether a long-only strategy designed for rising markets can be backtested using only historical periods that rose, while excluding declines and joining selected price segments into a synthetic series. Its central distinction is…
The note discusses how to identify market periods that favor trend-following or mean-reversion strategies. One suggested approach is regime modeling with Hidden Markov Models, with two research papers offered as starting points. The answer also cautions that…
The document explores a technical analysis workflow for finding significant swing highs and lows, then plotting trendlines through those points. The example code identifies local extrema by comparing each bar with its immediate neighbors. It filters those…
The discussion separates signal generation in a realistic backtest from statistical inference about predictive relationships. For price-based rules such as moving-average crossovers or Bollinger Bands, the answers advise using the original historical prices,…