Module 4 · Data, Limits & Exam
Data Quality & Latency
A heatmap is only as good as its depth data.
In this lesson you'll
- Understand the data a heatmap needs (Level 2, MBO)
- Account for latency
- Know why identical-looking heatmaps can differ
The heatmap is only as good as its data:
- Full depth (Level 2), not just top-of-book, is required to see the whole book.
- Iceberg / large-lot detection needs MBO (Market-by-Order) data with an untouched sequence of events. Aggregated depth (market-by-price) can't reveal individual orders or queue position.
- Platforms/feeds differ. Two identical-looking heatmaps can rest on very different data quality, which changes what you can trust.
Key takeaways
- Heatmaps need full Level-2 depth; icebergs need order-level MBO data.
- Aggregated depth can't reveal individual orders or queue position.
- Identical-looking heatmaps can rest on very different data quality.
- You see the book with latency — algos act faster than you can.
Check your understanding
Pick an answer to reveal instant feedback. Answer every question to complete the lesson.
Why that's off: Top-of-book hides the depth and the refills entirely.
Why that's off: A daily chart shows none of the order book.
Why that's off: It specifically needs high-quality MBO data.
Hidden refills are only visible with order-level (MBO) data. Aggregated market-by-price depth can't reveal individual orders, so it can't reliably detect icebergs.
Why that's off: Latency is a real, important constraint.
Why that's off: Latency is a delay, not a crystal ball.
Why that's off: Latency affects timing, not whether walls exist.
The depth picture reaches you with delay. A co-located algorithm reacts to changes faster than a retail trader, so resting liquidity can be acted on before you can respond.
Educational content only. Nothing here is financial or trading advice. Order-flow and options patterns shift probabilities; they never guarantee outcomes.