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.
9,660 documents
The document discusses the Corwin–Schultz estimator, which infers bid-ask spreads from high and low prices over consecutive days. Its derivation relies on a geometric Brownian motion assumption and a relationship between one-day and two-day variance,…
The question concerns whether a short-horizon foreign exchange forecasting model should use bid prices, ask prices, or a blended series, and how to evaluate trades in a backtest. The response distinguishes measuring market movement from simulating executable…
The discussion explains how market makers provide liquidity and manage the risks created by their quotes being executed. Their strategies consider whether incoming orders may contain information about future price moves, and they adjust quotes to control an…
The document considers software options for building a dynamic financial charting tool centered on candlestick plots. The person asking prefers Python because their programming experience is limited and would like interaction features such as zooming and…
The document considers whether countries or exchanges have banned algorithmic trading and whether such cases could support a comparison of markets. The responses report no known current exchange that operates only through traditional pit trading, making a…
The document explains why predictive accuracy alone does not establish that a trading strategy will earn money. Forecasts and backtests can omit execution costs, commissions, order-book depth, bid and ask prices, and the effect of the strategy’s own orders…
The document clarifies that VWAP serves both as a benchmark and as an execution tactic. As a benchmark, it is the day’s volume-weighted average traded price, with certain special trades excluded. As a tactic, a broker’s algorithm aims to match its share of…
The document compares several definitions of high-frequency trading and explains why trades per second alone do not settle the question. Some answers focus on very short holding periods or tick-level data, while another emphasizes latency-sensitive…
The document concerns retrieving historical interest-rate curve values through Bloomberg’s TIA Python wrapper. The curve field shown in the question returns a current value, and the accepted response says that the relevant ICVS curves are not available…
The document considers whether a developer with an existing FIX engine should implement a FAST protocol engine from scratch. The only answer advises against building a custom implementation, arguing that established commercial libraries have been tested and…
The document discusses how an order and position management component should work in automated trading. Its central recommendation is to maintain an internal view of orders, fills, positions, and risk rather than relying entirely on a broker or exchange.…
The document considers whether a stock market maker can profitably quote on an exchange when inventory arrives from customer trades or another desk at an unpredictable rate and size. The response treats inventory changes as information for the pricing model:…
The document collects public claims and opinions about Renaissance Technologies’ Medallion fund rather than establishing a definitive account of its trading strategy. Suggested techniques include algorithmic trading, information theory, hidden Markov models,…
The document describes a proposed research approach: define learning environments from market microstructure models with known solutions, then use reinforcement learning to see whether agents recover features of those equilibria. As an initial example, the…
The document discusses how to represent brokerage fees, bid-ask effects, and slippage when simulating daily stock trades without quote data. It argues that a single relative transaction-cost rate cannot be assumed realistic across all situations: costs vary…
The document discusses a worked example from research on price impact, where market sells consume bid depth and limit bid additions replenish it. The response interprets the sequence in ticks: selling 15 shares against depth of 5 removes three levels, while…
The document considers how to use rapidly changing limit order book features, such as spread and bid-ask imbalance, to predict a future midprice. It suggests testing recent history or aggregation and emphasizes that useful inputs may include time of day,…
The document discusses whether traders can forecast future option prices directly, then outlines volatility-based approaches. It describes gamma trading as selling options when implied volatility is expected to exceed realized volatility, with delta hedging…
The document explains the role of the underlying asset position in a variance swap replication strategy. A short variance swap can be hedged with a strip of options across strikes, often represented in practice by an at-the-money straddle together with…
The document discusses how to estimate execution slippage for a strategy trading many stocks with orders that are small relative to average daily volume. One proposed starting point is a cross-sectional regression using bid-ask spread and order size relative…
The document discusses how much decimal precision trading and portfolio software may need for equity prices and share quantities. It explains that quoting increments depend on the product and venue, and that rules governing displayed quotes do not…
The document considers how to infer whether a trade was buyer or seller initiated when a feed omits the aggressor-side field. A basic quote test labels trades at or above the best ask as buys and those at or below the best bid as sells, leaving trades inside…
The document asks whether the size at the best bid and ask in a limit order book can be modeled to predict when the current top price level will disappear. It clarifies that quote entries have size, while volume refers to executed trades. One suggested…
The document outlines three approaches to modeling limit order books. Zero-intelligence models represent event frequencies and order sizes with relatively simple distributions. Two-dimensional flow models track changes to the best bid and ask queues.…