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 answer clarifies that NASDAQ ITCH is itself a form of tick data, rather than a separate alternative to tick data. ITCH messages require the user to reconstruct the limit order book in order to track the best bid and ask for each stock, a process…
The document frames a numerical problem from quantitative finance: differentiating a discrete sum with respect to a continuous upper limit to estimate first- and second-order sensitivities. It gives an exponentially decaying summand with a fractional-power…
The document asks how a market maker might assign monetary value to an order resting at a particular price level and queue position, using BTC-USD as an example. It offers no valuation formula or empirical method. The response says that pricing queue…
The discussion explains why red-black trees are often proposed for limit order book price levels and why they are not universally best for production trading. They provide worst-case logarithmic insertion and deletion, which can help when price levels are…
The document asks for research on tick and quote filtering methods with few tunable parameters. Its answer proposes a short set of basic validity checks: reject quotes where the bid is at or above the ask, flag trades printed well outside the quoted spread,…
The document explains that unmatched buy and sell interest is organized in an order book. Level 1 quotes show the best bid and ask alongside completed trades, while Level 2 quotes expose additional price levels and their orders. A trader can generally view…
The document addresses how to store order book states reconstructed by replaying add, cancel, and trade messages. Saving a full serialized state after every message can consume substantial disk space, especially when many consecutive states are similar. The…
The document describes a data-cleaning approach for incorporating overnight prices into high-frequency stock-index data. The example method first stores the median overnight price as an observation at 23:00 on the day trading began, then changes that…
The document explains that the programming language used to develop a trading strategy does not have to match the language used to communicate with a broker. It describes a common architecture in which analysis is performed in one language and orders or…
This discussion explains why tick-to-trade latency depends on the full market-data and order-routing path, not just the speed of strategy code. Protocol choices, network distance, operating-system and kernel processing, packet parsing, and order transmission…
The document discusses a high-frequency machine-learning target designed to distinguish price changes large enough to overcome transaction costs from smaller mid-price movements. A rise in the mid-price alone may still lose money if a trader buys at the ask…
The document explains, at a simplified level, how an exchange may process market orders when buy and sell orders arrive close together. In the described model, market orders execute against resting limit orders on the opposite side of the book; market orders…
The document explains why equity orders do not always experience simple time priority and describes several sources of execution differences. Some venues use matching rules that depart from standard price-time priority, such as broker priority or parity…
The discussion considers whether high frequency trading contributed to the sharp market moves of August 2015. One cited analysis of participant data found that the share of high frequency traders in the best bid and offer remained broadly stable during…
The document defines a trade markout as the difference between a current trading price and a future reference price, often a midpoint or a more refined estimate of fair value. The future observation is taken after a chosen horizon, such as a number of…
The document considers how to estimate high-frequency trading activity for an individual security in near real time when trader identities are unavailable. Suggested proxies use order-book behavior: count quote cancellations to detect elevated order churn,…
The discussion surveys ways to estimate or forecast volatility from high-frequency market data. It points to realized volatility research based on tick observations and notes that high-low data can offer an alternative when using every tick is impractical. A…
The document considers the computational cost of delta-gamma value-at-risk when a portfolio has many risk factors. A full cross-gamma matrix can be large, making the calculation appear comparable in difficulty to simulation methods. The question asks whether…
The note explains how to simulate which order arrives next in an order book when each order type has an arrival rate. If order arrival times are independent exponential random variables, the probability that a particular order is first equals its rate…
The document discusses domain-specific ways to validate an order book reconstructed from NASDAQ ITCH messages. It recommends canned scenarios that exercise basic book behavior and edge cases, including checking that an execution message reports the trade…
The note addresses whether a trader can tell, from a decrease in displayed depth without a trade, if cancellations occurred ahead of or behind their order in an order-based limit order book. With aggregate depth data alone, the answer is to estimate the…
The document points readers seeking an open source liquidity seeking algorithm toward research on splitting orders across liquidity pools with a stochastic algorithm. It presents this paper as a way to learn the theoretical mechanics of liquidity seeking. It…
The document considers how low transaction costs can be for high-frequency trading, where small price movements make fees consequential. Its response says the answer depends on what counts as a fee, including whether regulatory charges are included alongside…
The document considers how to store intraday tick data and produce candles efficiently on very limited hardware. The original approach stores tick records in MySQL, adds indexed time buckets, and computes open, high, low, close, and interval boundaries…