Tick Data, Order Book Reconstruction, and Market Data Quality Limits
Summary
The article explains event-based tick data through changes to a limit order book: orders arrive, are canceled, or trade against resting quotes. With this event stream, a researcher can reconstruct the visible book, subject to venue rules and the depth included in the feed. It contrasts this with periodic snapshots, which report market state at intervals and therefore omit what happened between observations. The article also distinguishes snapshots, trade-by-trade records, and order records, noting that delivery can be batched rather than truly instantaneous.
It highlights engineering and research limits: high message rates, UDP packet loss, processing delays, incomplete market depth, and errors that can corrupt a reconstructed book. Vendor datasets may also contain gaps or inconsistent timestamps because collection systems sample changing files, suppress unchanged records, or suffer operational failures. The examples and timing claims describe particular market arrangements and may not generalize to every venue or current feed. Its main practical lesson is to validate and clean market data, since missing or faulty observations may not be recoverable from the dataset itself.
Key ideas
- Tick feeds can represent order arrivals, cancellations, and executions used to update an order book.
- Snapshot feeds show market state at intervals and leave activity between snapshots unseen.
- Feed depth and delivery timing vary by venue, and records may be batched rather than immediate.
- Packet loss, latency, collection practices, and operational faults can create gaps or corrupt data.
- Researchers should validate and clean data because some missing observations cannot be reconstructed.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.