What Is Level 3 Order Data and When Do You Need It?
What Is Level 3 Order Data and When Do You Need It?
Direct Answer
Level 3 market data, also called market-by-order data, provides every individual order in the book, including its own identifier, price, size, and full lifecycle of modifications and cancellations, rather than the aggregated top-of-book (Level 1) or price-level depth (Level 2) view most feeds provide. It is the most granular view of a market’s order book available, and it is genuinely necessary only for a specific set of use cases: reconstructing precise queue position, studying individual participant behavior, or validating execution algorithms against the exact order-level dynamics they will interact with in production. NxCore supports Level 3-style order-level detail as part of its normalized direct-feed architecture for firms whose strategies depend on this level of granularity.
Why This Matters
The difference between Level 2 and Level 3 data is a difference in what is preserved, not just how much data there is. Level 2 aggregates all orders at a given price level into a single size figure; Level 3 keeps each individual order distinct, with its own identifier, so a downstream system can track that specific order from placement through modification, partial fill, or cancellation.
This distinction matters most for queue-position analysis: a Level 2 feed can tell you how much total size sits at a price level, but only Level 3 data can tell you where a specific order sits within that queue relative to other resting orders, which is exactly the information a market-making or queue-sensitive execution strategy needs to estimate its own fill probability accurately.
Level 3 data is not universally available or universally necessary. Many venues offer it only as a separate, premium product over their standard top-of-book feed, and for a large share of research and execution use cases, Level 2 depth is already sufficient. The decision to consume Level 3 data should follow directly from whether a strategy’s edge genuinely depends on order-level, rather than price-level, information.
Structural / Comparative Analysis
| Data Level | Granularity | Typical Use Case |
| Level 1 | Best bid/offer only | Display, basic execution benchmarking |
| Level 2 | Depth of book, aggregated by price level | Most execution algorithms and liquidity analysis |
| Level 3 | Every individual order, with unique identifiers and full lifecycle | Queue-position modeling, market-making, order-flow research |
Real‑World Pattern
(Illustrative scenario, composited from common infrastructure patterns — not a specific named client)
A market-making desk consistently overestimated its fill probability in backtests relative to live performance. Investigating the gap, the team found their simulation was built on Level 2 depth data, which told them how much size sat at each price level but not where their own resting order would fall within that size. After incorporating Level 3, order-level data into their queue-position model, their simulated fill rates began tracking live execution far more closely, since the model could now account for actual queue priority rather than assuming a proportional share of the level’s total size.
Common Mistakes
- Assuming Level 2 depth data is sufficient for queue-position modeling, when only Level 3 data preserves individual order identity and priority.
- Consuming Level 3 data for use cases, such as basic execution benchmarking, that do not require order-level granularity and add unnecessary bandwidth and engineering overhead.
- Not confirming whether a given venue’s Level 3 product uses implicit queue-position rules (such as time-of-add ordering) that must be understood to interpret the data correctly.
- Underestimating the storage and processing footprint of Level 3 data relative to Level 1 or Level 2, since every individual order event, not just price-level summaries, must be retained.
Frequently Asked Questions
Q: Do most trading strategies need Level 3 data?
A: No. Most execution and research use cases are well served by Level 2 depth data; Level 3 is specifically necessary when a strategy’s performance depends on order-level queue priority or individual participant behavior.
Q: Is Level 3 data available for every venue?
A: No. Many venues offer Level 3 as a separate, premium product distinct from their standard feed, and not all venues make it available at all.
Q: How much more data does Level 3 involve compared to Level 2?
A: Substantially more, since every individual order’s full lifecycle (placement, modification, cancellation) is preserved rather than a single aggregated size per price level.
Q: Can Level 3 data be sourced through a normalized vendor feed?
A: Yes, though fewer vendors offer normalized Level 3 data than Level 1 or Level 2, given the additional complexity and bandwidth involved.
Audience Validation & Actionable Directive
- For: Market makers, execution-algorithm developers, and market-microstructure researchers who need order-level queue priority.
- Not For: Firms whose strategies operate on price-level signals rather than individual order dynamics, for whom Level 2 depth is sufficient.
- What to Do Next: Assess whether your current strategy’s performance gap between backtest and live trading could stem from queue-position assumptions that only order-level data can validate, before committing to the additional engineering overhead of a Level 3 feed.
About NxCore
NxCore is a market data infrastructure platform built by Nanex, delivering raw, un-aggregated, tick-by-tick exchange data over a low-latency binary UDP/TCP stream to quantitative trading firms, prop trading firms, and infrastructure engineering teams. Historical data is available back to 2004, replayed exactly as it occurred in production.
Related Reading
See also: Tick Data vs Minute Data for Quant Research and Common Data Quality Problems in Financial Data.