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 discussion examines whether electronic market making demands less quantitative skill than statistical arbitrage, and whether low-latency programming alone can support a trading business. Its central practical point is that fast infrastructure does not…
The document considers how to choose buy, sell, or hold actions from historical stock data while accounting for transaction costs, bid-ask spread, and slippage. It cautions that the problem cannot be judged by the apparent size of its search space alone: a…
The discussion compares general-purpose NoSQL systems with databases designed for financial time-series data, especially tick storage and retrieval. It explains that web-oriented stores often prioritize fast key lookups, row-level operations, and horizontal…
The document discusses how to approximate a sell execution price when backtesting with one-minute OHLC bars. It offers simple proxies: the bar close as a representative price, the low as a conservative scenario, or the mean of the four OHLC values. The…
This short discussion compares Python wrappers for Interactive Brokers’ Trader Workstation API. It explains that IbPy may be easy to use, but its limitations include lack of Python 3 support and compatibility with an older version of the broker’s API, which…
The document explains why butterfly spreads with different strike widths may not have prices proportional to their maximum payoffs. It offers an approximation that treats each butterfly’s payoff as a digital indicator for whether the underlying finishes…
The document clarifies basic assumptions for a simple order-matching simulator. It treats the quoted price as a per-unit limit: a seller sets the minimum price they will accept, while a buyer sets the maximum they will pay. A match is possible when the…
The document asks where to find historical order-level data, including orders and their final statuses or execution reports, to test a trading system that processes FIX orders through limit order books and auction or continuous sessions. It points to two…
The document explains stock-price changes across short and long horizons. For immediate price movements, it describes how market orders consume available liquidity and how limit orders state prices at which participants are willing to trade. Dealers and…
The discussion addresses how to estimate a trading strategy’s capacity using historical data and market depth. The question notes that standard broker market-impact tools may not capture strategies with complex entries and exits or those whose opportunities…
The document explains that FIX is a messaging protocol and does not define a standard field that directly reports the current floating profit or loss of a filled order. The suggested calculation uses execution-report details such as fill price and quantity,…
The document describes a QuantLib-Python issue encountered when changing a historical fixing that has already been added to a floating-rate index. Adding the same date again raises a duplicate-fixing error. Although the underlying API has a force-overwrite…
The document outlines common roles for field-programmable gate arrays in high-frequency trading. FPGAs can process large streams of market data with simple, repeated computations at low latency. They may handle feed parsing, compute analytics, perform risk…
The document collects historical suggestions for traders seeking broker APIs usable from .NET or C#. It describes Interactive Brokers’ API as requiring TWS or IB Gateway, while noting a community-written C# client. Other replies mention web-service or…
The document describes the inputs needed for periodic portfolio rebalancing: current weights, updated expected returns or alpha forecasts, forecast uncertainty, transaction costs, risk aversion, objective function, and portfolio constraints. Its central…
The document examines why an ETF’s midpoint may remain apart from a net asset value estimate computed from constituent midpoints. It emphasizes that midpoint comparisons do not establish an executable arbitrage: creation or redemption involves trading the…
The document describes ways traders may act on information before it is fully reflected in prices. A trader expecting an asset’s fair value to rise can buy while the market price remains below that estimate, then unwind as the market incorporates the…
The document explains two fields in a Gain foreign exchange tick dataset: an identifier at the start of each record and a final letter flag. The accepted response describes the identifier as a vendor-assigned unique tick ID that can distinguish multiple…
This discussion organizes risk management for high-frequency trading into infrastructure, model-specific, and system-wide controls. Infrastructure oversight covers software and hardware reliability, code review and testing, connectivity and reconnection,…
The document explains how a trading venue can price a match when a buyer is willing to pay more than a resting seller’s limit price. In the described order-book model, one participant first posts a limit order and provides liquidity; a later marketable order…
The document explains a reported Saturday loss of access to Interactive Brokers historical data and asks whether TWS or IB Gateway can be directed to another regional server. The answer describes changing the server peer setting in the application’s…
The document explains why simply counting buyers and sellers does not determine price direction. It reframes the question around how new limit orders affect activity in the order book, using a Queue Reactive model in which event probabilities depend on queue…
The document explains how trading contracts for difference can affect the price of an underlying asset indirectly. A CFD is a claim against its issuer, so a customer trade does not necessarily produce an immediate matching trade in the underlying market.…
The document raises several ways to strengthen systematic strategies: estimate portfolio risk using correlations instead of relying only on equal weighting or inverse-volatility weighting, model transaction costs, and add constraints to mean-variance…