A Point of Control is easy to draw and easy to over-interpret. It identifies the price bin that accumulated the most volume inside a defined profile window. It does not say that price must bounce there, break through it or treat it as support or resistance on the next visit. Those are outcomes, and the line alone contains no outcome data.
The useful question is therefore not simply, “Where is the POC?” It is: “When price returns to a completed profile’s POC, what counts as a revisit, what counts as acceptance, what counts as rejection, and how often did each event occur under the same rules?” Once those definitions are fixed, the chart can answer the question without hindsight.
What the Point of Control actually tells you
A volume profile divides a chosen time window into price bins and assigns volume to them. The Point of Control, or POC, is the bin with the greatest assigned volume. That makes it a summary of where activity concentrated during that window — not a forecast for the next one.
Two details matter before any retest can be measured. First, the profile window must be explicit: a daily POC and a weekly POC summarize different auctions. Second, the binning and volume-distribution rules must stay consistent. Change the anchor, resolution or data feed and the POC can move, so results from one configuration should not be presented as a universal market statistic.
A revisit is not the same event as acceptance
Price touching or trading through the prior POC establishes only one fact: the level was revisited. It does not tell you whether the market stayed there. Treating every touch as a “successful support test” combines three different states:
- Revisited: price traded back through the prior completed window’s POC.
- Accepted: after returning, price produced the required number of confirmed closes inside a defined zone around the node.
- Rejected: while armed, a completed bar traded through the node’s center and closed outside the zone; or, after testing had started with a close inside, a later completed bar closed outside before the required sequence was complete.
All of those transitions are evaluated only after the bar closes. If an armed node’s completed bar merely grazes the zone edge, never trades the center and closes outside, the node stays armed. It has not entered testing and it is not rejected. Calling that graze a rejection would inflate the count; calling it acceptance would do the same in the other direction. “No verdict yet” is a valid state.
Why a zone is more honest than a one-tick line
The POC is calculated from a price bin, so testing it against a single exact price creates false precision. A practical rule needs a tolerance zone that scales with both the profile’s bin height and current market movement. Otherwise the same one-tick tolerance behaves very differently on EURUSD, Bitcoin and gold, and a small change in feed precision can change the label.
The Volume Nodes & Value Migration implementation uses the node center plus a zone whose half-height is the larger of one profile bin or half the ATR recorded when the profile commits. Its default acceptance rule requires two confirmed closes inside that zone. These are explicit measurement choices, not claims that one tolerance is optimal for trading.
One chart still needs three separate denominators
The verified OANDA EURUSD 15-minute chart with Daily profile windows did not produce one catch-all “POC score.” Its panel kept separate samples for completed windows, resolved POC retests and resolved HVN retests. That separation matters because each row answers a different question. A revisit rate uses completed windows as its denominator; a POC acceptance rate uses resolved POC tests; an HVN rejection rate uses resolved HVN tests.
The POC-acceptance and HVN-rejection rows are therefore not complements and should not be added together. They refer to different node classes and different event samples. Results from that QA chart also belong to its symbol, timeframe, OANDA feed, Daily anchor and loaded history; they do not forecast the next retest.
This is the main lesson. A level can be revisited often without being accepted often. A high-volume node can also behave differently from the single highest-volume bin. Publishing the correct denominator beside every observed frequency makes those distinctions visible instead of hiding them in a headline. The linked free, open-source indicator exposes those separate counters for inspection on each chart.
Confirmed closes prevent hindsight labels
A forming candle can move into a POC zone, leave it and return before the close. If the classification updates on every tick, the same historical retest can appear accepted at one moment and rejected later. That is the same timing problem explained in what repainting actually is.
A reproducible retest test should update the state, counters and alerts only after a bar closes. In the implementation used for the measurements above, profile accumulation, node transitions and counters are confirmed-bar gated; resolved outcomes are latched. Bar Replay testing recorded more than 15 intrabar updates without changing an existing mark or counter, followed by one atomic update on the bar close. That verifies timing behavior, not trading performance.
POC migration answers a different question
A retest asks what happened when price returned to an old area of concentrated volume. POC migration asks where the concentration itself moved between completed windows. A higher next POC means the new window’s greatest assigned volume formed at a higher price; a lower POC means it formed lower. Neither direction guarantees continuation.
Keeping migration separate from retest outcomes avoids another common shortcut: assuming that a rising POC makes every old POC support, or that a falling POC makes every old POC resistance. Migration describes a change in the location of value. Acceptance and rejection describe what happened on a later return. The chart walkthrough shows both processes without treating either as a trade signal.
A checklist for measuring POC retests
- Fix the profile window. State whether the POC comes from a session, day, week or another anchor.
- Fix the construction. Keep bin resolution, volume allocation and data feed consistent for the sample.
- Define the zone. Use a tolerance that has a stated relationship to bin height or volatility.
- Separate touch from outcome. Count revisits, acceptance and rejection as distinct events.
- Wait for confirmed closes. Do not let a forming candle decide a historical label.
- Keep unresolved cases. Do not force every edge graze into a positive or negative bucket.
- Print the sample size. A percentage without its denominator is incomplete.
- Re-measure per chart. Symbol, timeframe, feed and regime can all change the observed frequency.
FAQ
Is the POC support or resistance?
Not by definition. It is the highest-volume bin in a completed profile. A later return can be accepted, rejected or remain unresolved; the outcome has to be measured separately.
Does a POC revisit imply a reversal?
No. “Revisited” records contact with the level. It says nothing about the direction or persistence of the move that follows.
Why not classify a wick through the POC immediately?
Because an intrabar probe can reverse before the close or continue through the zone. Confirmed-close rules give the historical and live chart the same decision point.
Is a volume node the same as a liquidity pool?
No. A volume node is derived from volume distributed across price bins. Liquidity pools are commonly inferred from swing levels, equal highs/lows or other resting-order structures. The distinction is expanded in the guide to liquidity sweeps and accepted breaks.
What if the feed has no volume?
The linked indicator reports that condition and falls back to equal weight per bar, which is a time-at-price approximation rather than a volume profile. Results from that fallback should not be described as volume measurements.