Order Flow & DOM Magnets
The claim: That reading the order book — big resting orders, 'magnets', absorption — lets you see moves before they happen.
~54% accuracy ceiling — entirely eaten by costs.
How it actually works
You watch Level-2 depth for abnormally large resting orders that price seems drawn toward, and for aggressive vs. passive order flow, to time entries.
We recorded real Level-2 data and built models on it directly — this was tested with genuine order-book data, not theory.
We tested it
The best directional accuracy plateaued around 54% — and that thin edge was completely consumed by spread and commission.
Order flow is a real, fascinating description of the present. As a directional edge for a retail trader, it did not survive costs.
Every figure comes from this project's own out-of-sample walk-forward testing on real costs — the full research record is in the Lab.
What it teaches you
Even genuine microstructure information isn't automatically tradeable. A 54% signal is worthless if the round-trip cost is bigger than the edge.
A perfect-looking pattern can still have no edge. The test, not the story, decides.
Put it to work — safely
Whatever you trade, the survivable part is the sizing. Use our free position-size & risk calculator to bound every trade before you take it, and read the live volatility state on the board.
Education, not advice. Nothing here is a recommendation to trade, or a claim about which direction a market will go — see Terms.
