What Is Institutional Order Flow? One Phrase, Two Things

Institutional order flow is the buying and selling of large participants: banks, funds and other firms whose orders are big enough to move a market. The phrase is used for two different things, though, and that is the part most pages skip. For some writers it means the order book that records those orders. For others it means marks on an ordinary candle chart that are read as the trace they left. The two need completely different data, and which one you mean decides whether you can test anything at all.

Which of the two does "order flow" mean?

It depends entirely on which page you landed on. Both meanings are in print, used confidently, with no warning that the other exists.

Family A, the order book. This group means depth of market, footprint charts, the tape and delta: records of actual orders and executions at each price. City Traders Imperium draws the line between two of those records cleanly. Its page gives liquidity as "pending limit orders in the DOM (not yet executed)" and volume as "actual completed trades (executed buy/sell orders)".

Family B, the candles. This group means structure read off a normal chart: order blocks, fair value gaps, liquidity sweeps, corrective candles. Writo Finance, writing from inside this family, states the split more directly than anyone else does: in the ICT framework, order flow is "not analyzed through a traditional order book or DOM (Depth of Market). Instead, traders can read it directly from the price chart using candlestick patterns." It goes on: "This definition sets ICT order flow apart from conventional order flow analysis, which relies on volume data and tick charts. The ICT approach is purely price action and time-based."

Source, and the date on the page	Which family	What it means by order flow
citytradersimperium.com, June 10, 2025, updated September 5, 2026	A, the book	DOM, footprint charts, volume profile, the tape
bookmap.com, September 5, 2025	A, the book	Level 2 data and the order book, from a vendor that sells a footprint product
theforexscalpers.com, July 16, 2026	A, the book	Bid and ask, the tape, delta and absorption, with a futures contract as its worked example
writofinance.com, May 23, 2026	B, the candles	Corrective candles against the trend, around a break of market structure
quantum-algo.com, April 14, 2026	B, the candles	Order blocks, fair value gaps and liquidity sweeps, read on TradingView
tradingfinder.com	B, the candles	Bullish and bearish order flow read structurally
blog.opofinance.com	B, the candles	Price action, market structure, order blocks, liquidity pools

Check which family a page belongs to before you take anything from it.

Can you actually see institutional order flow?

On a venue with a central order book, yes, you can see the book. On spot forex there is no such book to see.

City Traders Imperium sells order flow training and runs a prop firm. Its own page says: "spot forex has no central Order Book. It's a decentralised market", and therefore "This makes traditional order flow analysis in forex less reliable. You're seeing only a partial, synthetic view of the true flow of orders." On broker-supplied books it is blunter still: "Treat broker-side order books as sentiment tools, not institutional flow."

The same page sends readers who need precise order flow data towards currency futures, where real volume, delta and footprint tools exist. That is a different market from the one most retail forex and CFD traders are in, and worth knowing before you spend money on tools that assume it. ChartLabs carries forex, indices, commodities (metals and energy) and crypto. Futures is a market we do not carry, so nothing here is a route we can help you test.

So the honest answer to "how do I see institutional order flow" has three parts. On a futures venue you can read the real book. On spot forex or CFDs you are reading what one broker aggregates, which is a sample rather than the market. On a plain candle chart you are reading price and inferring the rest, which is a different activity that happens to share the name.

What does an ICT order-flow rule actually say?

One page states a rule precisely enough that two people could mark the same candle.

Writo Finance defines order flow as "the corrective candles that form against the primary trend direction, immediately before or after a break in market structure (trend continuation)".

That is a rule with parts you can argue about: what counts as the primary trend, what counts as a break of market structure, and how close "immediately" is. Pin those three down and you have something testable. That is a description of the rule, not an endorsement of it; the test is yours to run.

Can you compute delta or cumulative volume delta from ordinary chart data?

No, and the reason is a fact about data rather than an opinion about methods.

Delta is defined by which side of the quote each trade hit. The Forex Scalpers states it plainly: "Delta measures the difference between buying and selling volume at each price level. Positive delta means more aggressive buying (market orders hitting the ask), while negative delta means more aggressive selling (market orders hitting the bid)."

A candle records four prices and a volume figure for its interval. It does not record which of the trades inside it hit the ask and which hit the bid, because that distinction is thrown away the moment trades are summed into a bar. No arithmetic on candles brings it back. The same goes for a footprint, a DOM ladder and absorption: each of those is per price level order data, and a candle is a summary of a stretch of time.

ChartLabs charts data sourced from Dukascopy, tick-level where it is available, cleaned to remove anomalies and converted to 1-minute candles, plus the timeframes built from those, and one minute is the lowest interval we publish. Volume is there to read: the chart's legend always shows the bar's volume, and an optional Volume indicator can be added from the Indicators menu, where it opens in its own pane under price. It is off by default. What a volume figure cannot tell you, on any chart, is which side was the aggressor.

Quantum Algo asks the right question in its own heading, "How do you read order flow without a true order book?", and answers it with "proxies: cumulative volume delta, volume spikes at key levels, and the price reaction itself". The first of those cannot be built from ordinary candles at all, for the reason above. Anything labelled cumulative volume delta on a chart without side-attributed trades is an estimate produced by a rule the tool picked, and the tool should tell you which rule before you trade it. The page is at least honest that its proxies "will not match a professional footprint chart exactly".

What can you test, and what can you not?

A candle-derived rule can be written down and run over history, because everything it needs is in the bar. A book-derived rule cannot be tested on candles at all, however well you write it, because the data it reads was never recorded in them.

A trader with a real footprint feed on a futures venue is working with information that genuinely exists. A trader reading corrective candles is working with an inference. Both can be done carefully or carelessly.

If your rule mentions delta, absorption, the tape, resting liquidity or depth, stop: your data does not contain it, and a test built on a proxy is a test of the proxy. If your rule mentions candles, structure, gaps and levels, then everything it needs is present and you can measure it honestly.

How would you test an order-flow rule on your own history?

Write the rule using only what your candles actually hold, then be stricter about definitions than the source was.

1. Name the break of structure precisely enough that two people mark the same candle. "Price broke structure" is not a rule, it is a description. A rule says which swing point, measured how, broken by a wick or by a close.
2. Fix the lookback before you start. How far back does the primary trend get measured? Choose it once and write it down, because changing it afterwards to improve a result is how a test turns into a story.
3. Count how often the rule says nothing before you count anything else. A rule that fires twice a year is not worth testing further, and a rule that fires on every third candle is describing noise.
4. Only then record what the rule said against what happened next, with the chart stepped forward and the future hidden while you decide.

City Traders Imperium does carry a section headed "Back-Test Snapshot & Limitations", but what sits under it is a summary of academic research on market impact rather than a procedure, and its own caution is that meaningful analysis "requires massive datasets, and algorithms that detect hidden orders are prone to error, especially with large trades". The step-by-step part is still yours to build.

ChartLabs steps historical charts forward candle by candle, so you are deciding with the same information you would have had live, and every trade stores the chart exactly as you saw it when you took it. If you want the method before the market, start with our beginner's guide to backtesting a trading strategy, then the forex backtesting guide for session and pair choices, and backtesting accuracy, slippage and data quality for why costs decide whether a tested rule survives contact with a live spread.

How should you read a track record on an order-flow page?

By looking for what the figure is a figure of. One page on the subject advertises a headline strike rate under a "Public ledger" label, and its own small print undoes it.

Quantum Algo's note says the record "combines trade ideas posted publicly on TradingView before the outcome was known with premium signals recorded when they closed; it is not a brokerage statement, and past performance does not guarantee future results." Those are two different samples, selected in two different ways, added into one percentage. There is nothing there you could check or reproduce, and the page tells you so itself if you read far enough.

Not every number on the subject is like that. A page describing "Value area highs and lows where 70% of volume occurred" is giving the standard definition of a value area, not a claim about results. A page saying regulatory position data represents 70 to 90% of open interest is describing how much of a market its data covers. The test is always the same: does the number name an instrument, a period and a sample you could go and count yourself?

A broker's blog, in a section listing the method's weaknesses, says: "The presence of buyers or sellers at certain price levels doesn't guarantee future market direction."

Where to go from here

Decide which of the two meanings your rule depends on. If it needs the book, find out whether your market has one before you buy anything. If it needs candles, write it down properly and test it. You can run it bar by bar on cleaned data from Dukascopy with a free trial of ChartLabs: 7 days free, one plan at $20 a month or $200 a year, and no card needed to start.