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.
1,116 documents
The discussion names several approaches to forecasting asset volatility from returns. It mentions stochastic volatility and GARCH families, including FIGARCH and component GARCH variants for high-frequency settings, and highlights the HAR realized volatility…
The document discusses how to estimate variance when price observations arrive at uneven intervals. One approach is to aggregate log returns into daily returns and estimate daily variance from the squared daily returns, often using a zero-mean assumption.…
The document discusses ways to frame computational finance as a computer science master's thesis. Suggested directions include building investment recommendations from time-series regression models, evaluating predictive performance with machine learning on…
The document examines an apparent price discrepancy between two forex brokers: buying a currency pair at one venue and selling it at another. Although the offsetting positions can create a net-flat exposure to the underlying currency, the quoted spread is…
The post raises a question about deriving a Hamilton–Jacobi–Bellman equation for a high-frequency trading model with a limit order book. The model tracks cash wealth, stock price, and inventory: the stock price follows a diffusion, while inventory changes…
The discussion distinguishes modeling order book dynamics from implementing a matching engine. It introduces the Queue Reactive model as a way to simulate a limit order book: trade, insertion, and cancellation events at each price queue occur with…
The document asks whether high-frequency market-making systems commonly post liquidity in dark pools and lit exchanges at the same time, and whether dark-pool trading can affect inventory and displayed exchange prices. The response answers affirmatively and…
The discussion clarifies the basic counting units behind quote-to-trade statistics. A displayed quote consists of bid and ask prices together with their sizes. A change to any one of those four values can be counted as a quote change, so the measure depends…
The note proposes fitting an exponential order-arrival model, in which execution intensity declines with distance from the mid-price. Its two parameters control the baseline arrival level and how quickly the rate falls as quote distance increases. The…
The document explains several reasons a quantitative trading strategy may restrict itself to liquid, publicly traded assets that can be modeled. Low transaction costs matter when expected profits come from small pricing anomalies: trading costs can erase…
The document surveys ways to limit bid-ask bounce, which can create spurious short-term reversals in transaction-price data. One proposal is a size-weighted quote midpoint, intended to reduce the alternating bid and ask effect while supporting moving-average…
The document discusses how to prepare unevenly spaced quote data for technical analysis and other calculations that assume regular time intervals. Several quotes may arrive within the same second, while other seconds have no observations. One suggested…
The document describes a question about extending a GPU-oriented parallel binomial option-pricing method from European options to American options. In a standard binomial lattice, an American option's value at each node is the greater of its immediate…
The document addresses how to obtain one-minute intraday bars for a specified time window, rather than querying only by calendar date. It reports that Barchart offers an API with this capability, though the response characterizes it as paid and does not…
The document explains order flow imbalance (OFI) as a measure of changes in liquidity at the best bid and ask, and relates it to short-term price formation. It distinguishes liquidity provision, represented by changes in order-book queues, from liquidity…
The document explains a market-making perspective on high-frequency trading firms’ performance during the sharp US market decline on August 24, 2015. Market makers intermediate prices: they temporarily take on risk from participants seeking to trade and…
The document asks how to estimate realized volatility for individual time intervals rather than for an entire day. It describes realized variance at a chosen sampling frequency as the sum of squared log returns over that span. The example considers a daily…
The document explains why the microprice weights the best ask and bid prices using the displayed sizes at those levels. Those sizes carry information about available liquidity and the relative pressure in the top of the order book; replacing them with fixed…
The document asks how to predict near-term best bid or ask price direction from high-frequency limit order book data. One answer recommends reconstructing the actual book from a venue’s full order log when available, rather than relying on a simulated book.…
The document surveys quantitative and systematic approaches reported to have performed well during the 2008 crisis. It identifies managed futures, especially trend following in futures markets, as a major systematic category that did well. It also discusses…
The document explains why C++ skills can matter for quantitative researchers at high-frequency trading firms even when Python is more convenient for exploration, data analysis, and machine learning. It distinguishes research modeling from work that…
The document raises a practical implementation problem for a mean reversion strategy on multiple cointegrated assets. The cited model estimates an optimal inventory from regression estimates and their derivatives, but its output consists of fractional asset…
The document clarifies how pattern day trader buying power relates to maintenance margin excess. Its example distinguishes the account minimum from the maintenance excess: with no positions and an account balance of $26,000, the answer treats the full…
The document examines a market maker’s choice between keeping a limit order and replacing it. Replacing can lose time priority and reduce the chance of execution, while leaving an order in place can expose it to an unfavorable fill. The answers point to…