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.
4,613 documents
This notebook runs a previously selected US equities configuration on holdout predictions. The selected model, predictions, allocator, concentration, risk overlay, rebalance cadence, and transaction cost assumptions are inherited from earlier stages. Its…
This notebook demonstrates how a broker wrapper can reject unsafe orders using order-size and position limits, order-rate caps, asset restrictions, and a persistent kill switch. A synthetic broker makes the decisions observable: attempted orders are either…
This notebook connects position sizing with analysis of maximum adverse and favorable excursions. It demonstrates fixed-fractional sizing, where shares are determined by a portfolio risk budget and entry-to-stop distance, with a cap on position…
This notebook builds uncertainty intervals for ETF forward returns using Ridge regression and three conformal methods: split conformal prediction, conformalized quantile regression, and adaptive conformal inference. It evaluates them with walk-forward folds…
This notebook studies how allocation rules affect ETF strategies after a model has already ranked assets. It holds the prediction sets and rebalance calendar fixed, selects the same top-k basket for each comparison, and varies the weights assigned within…
This notebook tests whether portfolio construction can rescue a cost-defeated, every-bar NASDAQ-100 strategy. It selects leading baseline predictions using the declared selection universe, then runs the allocation comparisons on the full universe. The sweep…
This notebook takes leading validation prediction configurations from an equal-weight signal-stage baseline and tests alternative allocation methods on the same predictions. It compares score-based and conformal weighting with inverse volatility, risk…
This notebook compares stop-loss, trailing-stop, and time-exit overlays on fixed NASDAQ-100 microstructure backtests. The strategy makes decisions every fifteen minutes, while risk rules monitor prices every minute, allowing an overlay to exit a position…
This notebook compares split-conformal prediction, conformalized quantile regression, and adaptive conformal inference for forecasting ETF forward returns. In walk-forward folds, it trains on one subset, calibrates interval errors on another, and evaluates…
The document formulates pairs trading as a dynamic portfolio optimization problem. The spread between two related securities follows a Gaussian mean-reverting process, while its drift rate changes according to an unobserved finite-state continuous-time…
The paper applies generalized extreme value theory to tail returns in bitcoin futures markets to examine forced liquidation risk, leverage, and margin levels for both long and short positions. Its empirical analysis uses BitMEX perpetual futures. It reports…
The paper proposes a systematic investment approach intended to build wealth over a longer horizon through a sequence of short-term ETF purchase decisions. It combines evolutionary algorithms with a deep learning model to improve the timing of contributions…
This paper studies portfolio choice when an asset’s price trend follows an unobserved Ornstein–Uhlenbeck process. It derives optimal strategies for a trader with logarithmic utility under two information settings: one in which the trend is known and one in…
This document extends the Almgren–Chriss optimal execution setting, which assumes a fixed quantity to liquidate, to cases where the required trading volume is uncertain and becomes clearer as a delivery deadline approaches. Power markets are offered as an…
The paper studies portfolio liquidation games in which a trader exits the market once their position reaches zero. This absorption rule rules out round-trip trading and is framed as a no-statistical-arbitrage condition. In a model containing only sellers,…
The document proposes a systematic way to choose a stop-loss threshold by analyzing the distribution of maximum drawdowns. It frames stop selection as a decision guided by the goal of maximizing expected returns given available information, addressing the…
This study applies Twin-Delayed Deep Deterministic Policy Gradient (TD3), a continuous-action reinforcement learning method, to daily closing prices for Amazon stock and Bitcoin. The trading action represents both the position to take and the number of…
This paper describes a model-less convolutional neural network that takes historical prices for a set of assets and outputs portfolio weights. It is trained through reinforcement learning, with cumulative return serving as the reward. The training data cover…
The paper presents Deep Momentum Networks, which combine deep learning with the volatility-scaling framework used in time series momentum. A neural network learns trend estimates and position sizes together, with its parameters trained to optimize the…
This document examines how proportional transaction costs constrain leveraged investing in a market with one safe asset and one risky asset. It considers strategies designed to maximize long-term returns while maintaining an average volatility level, and…
This paper frames tail-risk management as an allocation across protections for distinct loss patterns: sudden crashes, volatility repricing and prolonged drawdowns. It develops a continuous-time conditional value-at-risk framework that combines long…
This study develops a day-ahead bidding and operating policy for a hybrid plant that combines wind generation with an electrolyzer. The policy sets power bids and schedules hydrogen production using contextual information, with linear decision rules learned…
The paper derives an optimal portfolio strategy for an investor maximizing expected logarithmic utility when multivariate stock-return drift is hidden and evolves as an Ornstein–Uhlenbeck process. The investor learns about drift from observed returns and…
The paper investigates whether trading position sizes can help control risk during periods of high volatility. It uses historical backtesting to examine how sizing models behave during crisis events, with particular attention to whether they can reduce value…