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.
21,023 documents
This post discusses portfolio risk measurement and argues that return volatility and Value at Risk do not capture losses beyond their usual thresholds. It proposes Expected Shortfall, also called ES, to measure extreme losses in the tail beyond a confidence…
This indicator displays the open, high, low, close, and time for a bar selected on the chart symbol. The user can choose a bar by clicking it, entering its index, or adjusting the displayed bar and timeframe with chart controls. The indicator can also…
TATR is presented as a simplified average true range measure focused on tick data. For standard bars, true range is the largest of the high–low range and the absolute gaps between the current high or low and the previous close. For tick data, the document…
The Open Oscillator tracks how the current open relates to the highest and lowest open prices in a chosen lookback range. It derives two values from those extremes and the current open, then smooths each with an exponential moving average to create signal…
This article argues that technical analysis should be treated as a probabilistic aid rather than a dependable forecast. It warns against trusting a single signal, applying one method in every market, and holding a position after price action invalidates the…
This document outlines an attempt to build a trend indicator using a neural network and Kaufman’s adaptive moving average as a reference. The described initialization phase trains the network on bars numbered 200 through 300 by selecting weight coefficients.…
This note proposes screening Chinese equities for a daily trading range above 1%, activity during 2021, and a measure described as institutional buying. The example defines the institutional condition as a rolling 20-session measure where at least 20% of…
The document distinguishes two broad uses of multi-factor models: regression models, which focus on explaining outcomes, and ranking models, which aim to identify stocks with attractive returns. Within the regression approach, it describes two ways to obtain…
This literature summary describes a stock-market return forecasting method that decomposes returns into dividend yield, profit growth, and price-to-earnings multiple growth. It says the source study uses the different time-series properties of these…
This report examines whether Bitcoin blockchain, exchange, investor, and miner measures move alongside price volatility in bull and bear phases. It groups indicators into institutional valuation and unrealized-profit measures, exchange liquidity and ETF…
This stock selection idea ranks shares by volume ratio and retains the top 100, then filters for stocks listed for more than 30 days whose 30-day moving average is rising. The intended rationale is to combine unusually high recent trading activity with a…
A stock screen for a moving average crossover can return no matches if its query loads only the target date. The note explains that time-series indicators need earlier observations to calculate their rolling values, so the data query should begin before the…
This code sample implements parts of the Alpha101 factor set using historical equity fields such as open, high, low, close, volume, returns, and volume-weighted average price. The formulas combine rolling ranks, moving averages, correlations, price changes,…
The ms-Candle Index estimates the direction and strength of each price bar using its range, body, and possible gaps. It is presented as a continuation of an earlier candle indicator, motivated by the limits of judging direction from open and close prices…
This document describes a variation of the Relative Strength Index that can apply a moving average to its input before calculating the indicator. The available smoothing choices are simple, exponential, smoothed, and weighted moving averages. Setting the…
This note presents a behavioral-finance explanation for a possible link between World Cup final outcomes and investor activity. A national team’s victory may lift local sentiment and trading, while defeat may discourage participation. The analysis compares…
The document explains how to decide whether to remove sector exposures from equity factor portfolios. It splits a stock characteristic into a within-sector component, which ranks companies against peers, and an across-sector component, which captures…
This A-share stock screen combines a metaverse sector classification, prior-day turnover above 8%, and a revenue comparison requiring 2021 revenue to exceed 2018 revenue by more than 10%. The rationale is to bring together sector exposure, trading activity,…
This document describes an indicator that adds a three-level trend-strength display to ColorJFatl. It uses the standard deviation of the indicator and two configurable thresholds to show colored dots on its moving average. A small dot signals a medium…
This overview surveys six ICLR 2022 studies: periodic time-series forecasting, model-based policy optimization, disentangled representations, Stein-based sampling, deployment-efficient reinforcement learning, and training with privileged observations. The…
This brief indicator description explains regression channels built around a fitted price trend. In the linear version, two parallel boundaries sit above and below the regression line. Their offset is set by the largest deviation of closing prices from that…
This article explains why selecting useful input features matters in predictive modeling. Irrelevant variables can add training time, contribute to overfitting, and make it harder for a model to generalize. Feature selection is presented as a way to reduce…
This indicator classifies candles by dividing the distance between each bar’s high and low into three equal sections. It then compares where the open and close fall within those sections to determine a bullish or bearish pattern. The resulting display is…
This document describes a historical analysis script intended to help day traders identify the times and days when a strategy performs best. It focuses on win-loss ratio and reward-risk ratio as measures for comparing strategy outcomes across trading…