Where should you anchor a VWAP?

To an event that a lot of participants can identify without hindsight: a session or week open, a gap, a scheduled release, the extreme of a completed swing. The test is whether you could have named the anchor before the move you now want to trade.

How is anchored VWAP different from session VWAP?

Only the starting point differs — the calculation is identical. Session VWAP resets on a clock, anchored VWAP resets on an event you choose. That makes anchored VWAP more expressive and much easier to abuse, because the clock cannot be tuned and your judgement can.

Can you anchor a VWAP to a volume level instead of a time?

Yes — anchoring to the moment a [point of control](/glossary/point-of-control/) or a high volume node was established is common. It is still an event anchor; the only requirement is that the event be identifiable at the time it happened rather than reconstructed later.

Glossary

Anchored VWAP

Also called: AVWAP · event VWAP

Anchored VWAP is a volume-weighted average price started from a chosen event — an open, a gap, a swing high, a news release — instead of the session boundary. It shows the average price paid by everyone who has traded since that event, which is a way of asking who is underwater. The anchor is the entire model.

What the line actually represents

Take a specific moment and count only what has traded since. The resulting line is the breakeven price of a cohort: every participant who entered after that event, in aggregate, is above water when price is on one side of it and below water on the other.

That reframes the reading. Ordinary VWAP answers "where was business done today". Anchored VWAP answers "who has control since the thing that mattered", and the interesting anchors are the ones that created a cohort — a gap that trapped a side, a release that repriced the instrument, a high that everyone can point to. The connection to trapped traders is direct: the anchored line is where their pain stops.

The anchor is the only parameter, and it is subjective

There is nothing else to tune. No length, no smoothing, no weighting choice. All the freedom sits in one decision, and that decision is unconstrained — any bar on the chart is a legal anchor. A tool with one free parameter and no rule for setting it is a tool that will fit anything you point it at.

The discipline is therefore procedural, not analytical. Anchor to events, never to outcomes: an anchor is defensible if you could have named it before the move happened — session open, gap edge, scheduled release, the extreme of a completed swing. If you had to look at today's price to decide where to anchor, you have drawn a conclusion, not a level.

Using more than one

Several anchors on one chart are legitimate — a session anchor, a weekly anchor, an event anchor — and they are informative when they converge, because convergence means several cohorts share a breakeven. Their disagreement is equally informative, and usually means there is no shared reference to trade against.

The cost is that each additional line raises the chance that something is always near price. Cap the number in advance, keep the same set every day, and treat a line you added this morning as evidence of nothing.

A worked example

In a synthetic ES sequence, a scheduled release at 14:30 prints a 12-point range in ninety seconds and the instrument settles 9 points higher. Anchor a VWAP at the first tick after the release: it starts at 5 340.50 and, after two hours of two-sided trade, sits at 5 344.75 while price is at 5 349.

Two days later price returns to 5 345.00 on falling volume and the anchored line has drifted to 5 345.25 — a 0.25-point separation. Everyone positioned since the release is, on average, flat. That is a testable statement with a defined invalidation: below the line, the cohort is losing. Compare it to the session VWAP that day at 5 351.00 — nearly six points away and describing a completely different population.

The trap

Anchor shopping. You scroll back, try eleven starting points, and keep the one whose line touches the current price. Such an anchor always exists, on every chart, in both directions. The tool cannot tell you that you found it by searching — the line looks exactly as clean as one chosen in advance.

Retest rates get quoted for anchored VWAP, and the figure moves with whoever is quoting it. None of the versions says what counts as a "major" swing, what counts as a "retest", on which instrument, over which window — so none of them is falsifiable as stated, and the disagreement between them is the more useful signal. The fix is mechanical, not statistical: write down the anchor and the reason before the setup exists, keep the log including the anchors that led nowhere, and count them all. An anchored VWAP you can only justify after the fact is a drawing, not a measurement.

Frequently asked

Where should you anchor a VWAP?: To an event that a lot of participants can identify without hindsight: a session or week open, a gap, a scheduled release, the extreme of a completed swing. The test is whether you could have named the anchor before the move you now want to trade.
How is anchored VWAP different from session VWAP?: Only the starting point differs — the calculation is identical. Session VWAP resets on a clock, anchored VWAP resets on an event you choose. That makes anchored VWAP more expressive and much easier to abuse, because the clock cannot be tuned and your judgement can.
Can you anchor a VWAP to a volume level instead of a time?: Yes — anchoring to the moment a point of control or a high volume node was established is common. It is still an event anchor; the only requirement is that the event be identifiable at the time it happened rather than reconstructed later.