Learn order flow
Sixteen articles on the mechanics that order flow records in CME futures, with the examples centred on NQ.
They are complete: there is no reduced version here and a fuller one elsewhere, and no form in the way. Every schematic is original.
Where to start
The order below has a reason. The first article sets out the mechanics, what moves price and what gets recorded, and without that base the rest reads like a collection of patterns. The second explains the tool everything else is read on, and the next four cover the core concepts built on it.
The other ten each take a single narrower question and follow it to the end, and they can be read in any order. They start with what the record contains and what it leaves out, move on to calibration per contract and where a pattern sits in a move, and end with the decisions the reader brings: which swings to compare, what counts as an unfinished auction, which price counts as the close, where yesterday's session ends, when a level counts as revisited and which candle counts as the largest-delta one.
- 01 Start here
What order flow is and what it records
Who moves price, why the flow is a record of what has already happened, how the DOM, the tape, the footprint and the profile fit together, and what the record leaves out.
- 02 The tool
How to read a footprint chart
What each column means, how the record is built, and why the diagonal comparison and the horizontal one give different answers from the same cells.
- 03 The concepts
Absorption vs exhaustion: telling them apart
Effort against result, how absorption differs from a void and from the two kinds of exhaustion, why delta cannot separate them, and why a well-identified absorption can still mean nothing.
- 04 The concepts
Footprint imbalances: when they mean something
The two different concepts that share the name, why the 3:1 threshold is a setting, and what changes when several imbalances stack.
- 05 The concepts
Delta and CVD: what they measure
Why delta is not direction, what accumulating it adds, why the CVD level depends on where the count starts, and what a divergence needs before it says anything.
- 06 The context
Volume profile: POC, value area and its limits
What the histogram represents, why the 70% value area is an inherited convention, why the POC depends on the window, and what the profile discards by design.
- 07 The data
Bid/ask footprints vs tick-direction data
What a footprint needs to know to attribute a trade, what changes when that attribution is inferred rather than read, and how to tell which version you are looking at.
- 08 The scope
What a footprint chart cannot show
Resting liquidity that was never hit, quotes that were pulled, spoofing, iceberg size, and intent. Why the gap is structural, and why no paid feed recovers it.
- 09 The calibration
Why NQ imbalances do not read like ES imbalances
The same 0.25 tick on a much higher index level, and fewer contracts spread across more price levels. What that does to the 3:1 ratio and to stacked imbalances.
- 10 The location
Stacked imbalances: initiation or climax
The same three-level stack can start a move or end it. Why sources give opposite instructions, what the stack's place in the leg and its delta history add, and what the retest shows.
- 11 The reading
Delta divergence depends on which swings you compare
Declaring a divergence requires picking two highs, and which highs exist is itself a threshold setting. Why divergences stack through trends, and what makes the reading defensible.
- 12 The definition
Unfinished auctions: three definitions, one cell
Sources describe the same extreme cell as finished and unfinished. What the cell records, why tick-rule data empties it, and what row size and thresholds change.
- 13 The reference
Settlement is not the close
How CME prices the NQ daily settlement, why sources quote different time windows, which price a chart calls the close, and why an opening gap depends on the reference.
- 14 The cut-off
Where the regular session ends, and the prior-day high
CME publishes no regular session for NQ, and sources end it at 15:00, 15:15 or 16:00 CT. Where the 15:15 figure comes from, and how each choice moves the prior-day high and low.
- 15 The level
Naked POC: what counts as a fill
A prior session's POC stays naked until price returns, but a return is defined several ways. Volume or time, any session or regular hours, touch or acceptance: why the same record gives different lists.
- 16 The session
Trend day or failed breakout: where the delta peaks
A high-delta breakout candle can start a trend day or end the move. Why the candle alone cannot tell, what the next candles show, and why the largest-delta candle depends on the interval.
One idea runs through them
Order flow tools present their settings as findings. The 3:1 imbalance ratio, the 70% that bounds a value area, the deviation threshold that decides which swings exist, the minimum that makes an extreme cell count as an unfinished auction, the price taken as yesterday's close, the time at which yesterday's regular session ends, the candle length that decides which candle carries the largest delta, the distance at which a move counts as extended, the touch that retires a naked POC: all nine are choices a reader makes, and all nine are routinely described as though the market supplied them.
The tools keep their value, and shared conventions are useful precisely because they are shared. The criteria simply have to be fixed before the session, because with enough adjustable parameters you can always find the signal you already wanted.
What you will not find here
There are no entry signals or performance claims here, and no assertions about institutional behaviour that the record cannot support. Each article includes a section on the limits and false positives of the concept it explains, because that part is as necessary as the rest.
The charts are schematics drawn to explain mechanics. They show why something happens, and they are neither screenshots nor documentary evidence of live trades.