Anchored VWAP: where you anchor is the whole setup

You anchored a VWAP to the swing low, price respected it twice, and for about ninety seconds you felt like you had found something real. Then you dragged the anchor two bars to the left and the entire line moved somewhere else. That second

You anchored a VWAP to the swing low, price respected it twice, and for about ninety seconds you felt like you had found something real.

Then you dragged the anchor two bars to the left and the entire line moved somewhere else.

That second part never makes it into the screenshot. Anchored VWAP is the volume weighted average price calculated from a bar you choose rather than from the session open, and the choosing is doing far more work than anyone admits.

Most explainers treat the anchor as a detail and spend their word count on what to do when price touches the line. That is backwards. Anchored VWAP has exactly one input that matters, you select it by hand after you already know what price did, and that makes it the easiest tool on the platform to fool yourself with.

Anchored VWAP has one input and it is not in the settings menu

Open the settings on TradingView's anchored VWAP drawing tool and you can change the colour and the thickness. That is the menu. Everything with any influence on where the line sits is fixed by the arithmetic.

The arithmetic is not complicated. TradingView publishes it in five steps: take the typical price of each bar as (High + Low + Close) divided by three, multiply it by that bar's volume, keep a running total of both that product and the volume, then divide one by the other.

In plain English, it is the average price paid since the anchor, with heavier-volume bars pulling harder. No smoothing constant, no period length, no source dropdown worth arguing about.

So there is one decision in the whole tool: which bar counts as bar one. Every trader who has ever tweaked an RSI from 14 to 9 and called it optimisation should recognise what that means. You have a single free parameter and it takes 500 values on a 500-bar chart.

Standard VWAP resets on a calendar, anchored VWAP resets on your click

The distinction that matters is not visual, it is who decides the reset.

StockCharts puts the traditional version plainly: the calculation starts with the first bar of the day and ends with the last bar of the day. With the anchored version, the chartist chooses the first bar and the last bar is always the most recent one available.

TradingView's built-in VWAP ships ten anchor values, from Session through Week, Month, Quarter, Year, Decade and Century, plus resets on Earnings, Dividends and Splits. Every one of those is mechanical. A calendar decides, or a corporate action decides, and your opinion never enters the calculation.

The VWAP Auto Anchored indicator goes further, offering fourteen anchor periods that add Auto, Highest High, Lowest Low and Highest Volume to that list. Those last three are the interesting ones, because they are exactly the "significant bar" anchors traders reach for by hand, expressed as something a machine can find without you.

The platform already hands you a set of anchors that were not chosen with hindsight. The moment you drag one by hand, you have left that set, and you should know that you have.

Choosing the anchor after the move is hindsight wearing a lab coat

Here is the mechanism that quietly ruins most anchored VWAP work.

You load the chart. You can already see the rally, the pullback, and where price sits right now. Then you go looking for the anchor that produces a line price has respected. You find one, because on a chart with hundreds of candidate bars at least one anchor produces a line with three clean touches.

That is the archery version of painting the target around the arrow. The line looks predictive because it was selected for looking predictive, using data that did not exist when the anchor bar printed.

Nothing about the tool warns you, either. The line renders with exactly the same confidence whether you anchored to a rule or to a hunch, which is precisely why the discipline has to come from you. This is the same free-parameter problem that shows up when timeframe combinations get selected after the fact, and it fails the same way: beautiful in review, unrepeatable in real time.

An anchored VWAP drawn on a chart you have already read is not analysis. It is a very well-dressed opinion about a move that already finished.

— Internal note on indicator selection, Tradoki desk

A defensible anchor is one you could have named before the bar printed

The test is a single question, and it is worth writing on a sticky note: could you have specified this anchor a month ago, in writing, without knowing what price did next?

Anchors that pass are the ones defined by an event or a computable extreme. The highest-volume bar of the last quarter. The gap bar on an earnings release. The session open following a scheduled economic print. The lowest low of a fixed lookback. Each of those is findable by a script that has no idea what happens afterwards.

Anchors that fail are the ones defined by their own result. "The bar where the trend started" sounds objective until you try to code it and discover the definition is "the bar we now know was the bottom." Same for the swing low that happens to sit where the line works.

StockCharts gives the reasoning behind event anchors clearly: these events signal a change in market psychology, and price action before the change should be excluded because it does not reflect the same psychology. That is a sound reason, but it is not yet a rule. It becomes one when you name the event class in advance and commit to anchoring to every instance of it, including the ones that go on to embarrass you.

On forex the weighting is tick volume, which is a different measurement

This one catches people who moved a stock-market habit onto a currency pair, and it is a genuine data problem rather than a nuance.

VWAP needs volume by definition. Spot forex has no consolidated tape, so there is no authoritative traded-volume figure to feed it. TradingView states directly that forex pairs and CFDs from certain feeds lack trade volume information, and that where trade volume is unavailable, tick volume is used instead, indicating the number of price updates.

The volume profile documentation says the same thing in one line: tick volume for indices, forex and crypto CFDs, indicating the number of price updates.

So your EURUSD anchored VWAP is not weighted by how much got traded. It is weighted by how often that feed updated its price. Those two things correlate during normal conditions and come apart exactly when you care most, in thin sessions and around releases.

The practical consequence is that two traders can anchor to the identical bar on the identical pair and get two different lines, because their brokers ticked differently. On a centralised futures or equity market that does not happen. This is another place where the wrapper you trade through quietly changes the thing you are measuring, and it is the same data caveat that applies to volume profile work on decentralised markets.

Test the anchor rule, not the line

You cannot meaningfully test "anchored VWAP" because it is not a strategy. It is a reference line. What can be tested is a fully specified rule: this anchor definition, this interaction with price, this invalidation, this exit.

Once the anchor is mechanical, the rest is ordinary research work. Script the anchor, generate the line across history, and count outcomes on a sample large enough to mean something, which is considerably larger than the twenty or thirty trades most traders stop at. Then treat the results with the usual suspicion, because the strategy tester's default assumptions flatter almost everything.

One replication caveat worth knowing before you start: TradingView notes that VWAP Auto Anchored is not written in Pine, so there is no source code to examine. If you want the exact behaviour of that indicator inside a testable script, you are writing your own version and accepting that small differences in anchor detection will produce different lines.

The honest summary is that anchored VWAP is a good tool with one dangerous property. It gives you a clean, volume-aware reference for "what has this move actually cost people since the thing that started it," which is genuinely useful context. It also hands you a dial that can be turned until the chart agrees with you, after the fact, with no warning light. Decide the anchor rule while the right-hand side of the chart is still blank, and the tool behaves. Decide it afterwards and you are grading your own homework.

● FAQ

What is the difference between VWAP and anchored VWAP?

The calculation is identical and only the starting bar changes. Standard VWAP resets on a schedule, and TradingView's built-in version exposes ten anchor values running from Session through Century plus Earnings, Dividends and Splits. Anchored VWAP starts from a bar you select yourself and runs to the most recent bar on the chart.

Where should you anchor a VWAP?

The workable test is whether you could have named the anchor before the move happened. The highest-volume bar of the quarter, an earnings gap, or the session open after a scheduled release are all computable in advance. The low that happens to make the line look good is not an anchor rule, it is a search result.

Does anchored VWAP work on forex?

It plots, but it is not weighting what most traders assume. TradingView documents that indices, forex and crypto CFDs use tick volume, which counts price updates rather than contracts traded. A forex anchored VWAP is therefore weighted by update frequency on one data feed, so a different broker feed can produce a different line from the same anchor bar.

Is anchored VWAP a leading or a lagging indicator?

Lagging, and it cannot be anything else. Every value is a cumulative average of bars that have already closed. It describes what has been paid since the anchor, which is useful context rather than a forecast of the next bar.

Can you backtest anchored VWAP?

Not as an indicator, only as a fully specified anchor rule paired with entry and exit definitions. TradingView also notes that its VWAP Auto Anchored indicator is not written in Pine and its source cannot be examined, so exact replication means writing your own version.