Anchored VWAP Tutorial

What is an Anchored VWAP?

An anchored VWAP is the average price people have paid for a stock since a particular date in time.

VWAP in plain English

VWAP stands for Volume-Weighted Average Price. It is simply the average price of a stock, but it gives more weight to days when more shares traded. A day with 90 million shares traded counts much more than a quiet day with only 9 million shares.

A normal moving average treats every day the same. VWAP does not. It cares about how much actual money changed hands. The number it gives you is closer to what the average person who bought or sold really paid.

Why the anchor matters

The VWAP most traders see on their charts resets every morning when the market opens. That version only works for one day.

An anchored VWAP never resets. You pick one specific day to start from. From that day until now, it keeps calculating the average price without starting over. The result is a price level tied to a real event in the stock's past, like a major high, a major low, an earnings day, or the start of the year.

Why anyone would watch it

There are two simple reasons.

First, it shows the break-even level for the average buyer.
If the stock is trading above the VWAP that starts from last quarter's high, then the average person who bought since that high is currently in profit. If the stock is below it, the average buyer is sitting on a loss. People trade differently when they are up versus when they are underwater.

Second, big institutions care about it.
Large funds often buy or sell over many hours or days. Their performance is judged by whether they beat the VWAP. Because of that, the level becomes a reference point for the biggest players in the market. This is why price often slows down or turns around when it reaches an anchored VWAP.

Neither reason predicts the future. The line only tells you where the average buyer stands right now. What happens next is still up to the market.

The Math Behind the Line

Three decisions define the calculation: which price each bar contributes, how much weight it carries, and how far back the sum runs.

Step one: the typical price

Each daily bar contributes its typical price rather than its close:

A single closing print is one moment out of 23,400 seconds of regular trading. The average of the high, the low and the close is a better stand-in for where the day's business actually happened, and it needs nothing beyond the daily bar to compute.

Step two: weight by volume

Every typical price is multiplied by that session's volume, and the running total is divided by the running total of volume:

A worked example

Two sessions since the anchor:

• Session 1: typical price $100 on 1 million shares.
• Session 2: typical price $94 on 3 million shares.

The plain average of the two days is $97.00. The volume-weighted average is:

The heavy down session pulled the line $1.50 below where a simple average would have put it, because three times as many shares traded there. This is the whole point of weighting. The line follows where the money went.

Step three: the line gets heavier over time

Because both sums accumulate from the anchor and never reset, a new session moves the line by roughly its share of total volume. Day two can shift it a long way. Day two hundred barely nudges it.

Real numbers from the tool: with SPY anchored at its 52-week high, the very first plotted point equals that one bar's typical price of $758.91. Thirty-four sessions later the line sat at $743.76, and by then it took an enormous session to move it even a few cents.

The Deviation Bands

The shaded envelope around the line is a dispersion ruler. It answers one question: how far is this price relative to how far prices have spread since the anchor.

How the bands are built

The tool computes a volume-weighted variance over exactly the same window as the VWAP itself:

On the chart the inner ±1σ envelope is shaded more strongly, the outer ±2σ envelope more faintly, and both edges are dashed. The right edge is tagged so you can read which line is which without hovering.

The bands start at zero width

On the anchor session there is only one bar, so there is nothing to disperse. Variance is zero, sigma is zero, and all four band lines sit exactly on top of the VWAP. They fan open from that single point as sessions accumulate.

The practical consequence: for the first handful of sessions after any anchor, the bands are far too tight to mean anything. A price sitting outside ±2σ three days after an earnings anchor is describing a sample of three bars, which says nothing about how extreme the move is.

These are not Bollinger Bands

They look similar and behave very differently.

That last row matters most. A Bollinger band squeezes when the market goes quiet because old bars roll out of the window. Anchored VWAP bands have no window to roll out of. Once a violent stretch is inside the sum, it stays inside forever, so the envelope tends to widen and then hold. A quiet month after a violent one will not pull these bands back in.

Band width tells you about the anchor as much as the stock

Measured on SPY, the same stock on the same day, with different anchors:

A move that is "2 sigma" on a two-week anchor and a move that is "2 sigma" on a one-year anchor are not remotely the same distance in dollars. Always read the band against the anchor that produced it.

The Nine Anchor Points

The anchor is the only real input to this tool, and each preset has a precise rule. Knowing the rule is what stops you misreading the level.

Two quirks worth knowing

Quarter-to-date and month-to-date collide four times a year. In January, April, July and October the quarter and the month begin on the same session, so both tiles show the same anchor date and the same VWAP until the following month starts. Measured on SPY in July 2026, both read $748.14 from a July 1 anchor. Nothing is broken. The calendar simply lines up.

Year-to-date and past 1 year are different animals. Year-to-date is pinned to a calendar boundary and grows longer every day until it resets in January. Past 1 year is a fixed 252-session step that slides forward daily. In early January they are as far apart as they ever get: one is a day old, the other is a full year deep.

Reading the Screen

Every element on the page is either an input, a level, or a piece of context about the level. Here is the full inventory, laid out the way the screen is.

1. Parameters panel. The nine anchors, stacked. The active one is filled.
2. Ticker and live quote. The symbol, its price, and the move on the day. The dot pulses green when price is up, red when down.
3. Ticker search. Switches symbols.
4. The eight tiles. Every preset anchor's current VWAP and the date it resolved to, with the distance from price to that level: a green up arrow when price is above it, a red down arrow when below.
5. The chart. Price as the dark line, anchored VWAP as the orange one, with the shaded ±1σ and ±2σ envelopes around it.
6. Right-edge pills and tags. Current price on the dark pill, current VWAP on the orange one, with each band edge labeled beside them.

What the screen does

The figure covers where everything sits. The rest of the screen is behavior:

• Clicking a tile switches the chart to that anchor and leaves that tile highlighted. The panel on the left does the same job.
• Custom Date reveals a date picker under the panel, already bounded to the sessions the page has loaded.
• Hovering the plot gives a crosshair, a dot on each line, and a tooltip with that session's date, close, anchored VWAP and ±1σ range.
• The two pills slide apart when price and the VWAP nearly touch, so neither number gets buried.
• The line under the chart names the anchor, its date, and how many sessions have accumulated since. That session count is what tells you whether the bands have had time to open up.

Live price and daily bars

The header quote, the tile distances and the price pill all read a live price that refreshes every 20 seconds while the regular session is open. The plotted price line, the VWAP, the bands and the tooltip are all built from daily bars.

That is why, mid-session, the price pill can sit slightly away from the end of the plotted line. The pill is current to the last 20 seconds. The line is current to the last settled bar. The two converge once the day closes.

Tickers, links and access

• The URL carries both the symbol and the anchor, so a view can be shared or bookmarked exactly as you left it.
• The tool remembers the last anchor you used for the rest of your browsing session.
• Daily bars are cached per symbol for the calendar day, so revisiting a name is instant and does not re-pull history.
• SPY and SPX are open to everyone. Other symbols are a member feature.

The schematic above is the map. To see it on the real screen, open the tool with its guided walkthrough running: it steps through the anchor panel, the tiles, the chart and the ticker search on live data.

How Traders Use It

Nine ways this tool gets used, each tied to the anchor that answers the question.

1. The break-even line after a top

Anchor at the 52-week high. Everyone who bought since the peak has, on average, paid this price. While price is underneath it, every rally is running into people who are finally getting back to flat and would like out. That supply is the reason the level so often caps the first attempt back.

2. The floor from a capitulation low

Anchor at the 52-week low. Buyers from the bottom are collectively in profit above this line. When a pullback reaches it, those buyers are back at cost, which is exactly where a lot of them decide to defend or add. It is one of the more reliable places for a pullback to find a bid in an uptrend.

3. The post-earnings scorecard

Anchor at Last Earnings. This is the cleanest answer to "is the market still paying up for that report?" Above the line, buyers since the print are winning and the reaction is holding. Below it, the reaction has been given back and the print is being repriced. It also resets every quarter, which keeps it relevant in a way a fixed level never is.

4. Benchmarking the calendar

Anchor year-to-date, quarter-to-date or month-to-date. These are the periods portfolios are reported on. The year-to-date VWAP is roughly the average cost of everyone who has traded the name this year, which makes it the line that decides whether the average holder is showing a gain in their statement.

5. The institutional cost basis

Anchor at the Highest Volume Day. That session is where the most shares changed hands in a year, which usually means it is where the largest repositioning happened. Anchoring there measures the average price paid since the market's biggest day of business in the name.

6. Event anchors

Use Custom Date for the moment that actually mattered in the stock: a guidance cut, a product launch, an index addition, a Fed decision, a merger announcement, the day a gap opened. If you can name the day the regime changed, you can measure everything that has happened since from the right starting point.

7. Confluence between anchors

Use the tiles. When two or three anchors resolve to nearly the same dollar level, that price is doing several jobs at once: it is the break-even for the top buyers, the cost basis since earnings, and the year's average all at the same time. Clustered levels carry more weight than any single anchor.

8. Measuring how stretched a move is

Once an anchor has enough sessions behind it, price riding the outer band is running well beyond what the average participant has paid. It does not mean a reversal is due. It means the move has extended relative to its own history, which is useful for sizing, for taking partial profits, and for deciding not to chase.

9. Placing risk

The line and the band edges give you a level that updates itself. A thesis built on "buyers since the last report are in control" is invalid the moment price closes decisively below the earnings-anchored VWAP. That is a cleaner definition of wrong than a round number picked off the chart.

The Theory: What the Bands Do and Do Not Say

This chapter is the one that keeps you out of trouble. The bands carry a sigma label, and that label invites a claim the math does not support.

The claim people assume

A standard-deviation band on a chart suggests the familiar rule: about 68% of observations inside one sigma, about 95% inside two. That rule describes independent draws from a normal distribution. A price path is neither independent nor stationary. Today's price is mostly yesterday's price, and the whole run drifts.

What the measurement actually returns

Here is the same stock on the same day, measured across different anchors. The last two columns count how many of the closes since the anchor actually finished inside each band.

No row lands on 68/95, and the rows do not agree with each other. The year-to-date anchor kept only about a fifth of its closes inside one sigma. The short anchor kept more than half.

Why the numbers come out that way

Sigma here measures how widely the typical price has spread around the volume-weighted mean of the entire run. In a trending market that spread is dominated by the trend itself. Price starts at one end of the range, ends at the other, and spends comparatively little time near the middle where the average lives. The longer and more directional the run, the smaller the fraction of time price spends inside one sigma.

A short, choppy anchor gives the opposite result. Price oscillates around a mean that has barely moved, so containment climbs.

Why the level has pull at all

Three mechanisms, none of them mystical:

• Reference pricing. Execution desks are benchmarked against VWAP. A level that grades performance becomes a level that attracts orders.
• Break-even behavior. People manage positions around their own cost. An anchored VWAP is a decent proxy for the crowd's cost since a memorable event.
• Reflexivity. Enough participants watch the same line that orders cluster there, which makes reactions at the line more likely, which keeps people watching it.

All three are behavioral. None of them guarantee anything on any given touch, and all three weaken when the anchor is one nobody else would have chosen.

An anchored VWAP measures the crowd's cost. It never promises what the crowd will do about it.