VWAP

The one volume tool professionals actually watch

Every lesson in this unit so far — volume confirmation, OBV, A/D Line — came from the retail technical-analysis tradition. VWAP is different. It was born on institutional execution desks and migrated to retail charts later. Kaufman:

Executions can be placed as a VWAP order, a volume-weighted average price… It is most convenient for hedge funds that do not want to force prices higher by placing an excessively large buy order at one point during the day, or even on the close. Equally spaced orders, which ignore liquidity, use TWAP, a time-weighted average price. — Perry Kaufman, Trading Systems and Methods

And:

Institutions with large orders must trade during times when the volume is heavy in order to keep slippage under control, or they may take advantage of VWAP orders… where partial orders are automatically fed into the market based on an algorithm that balances time intervals and volume.

The key insight: VWAP started as an execution algorithm, not a chart indicator. A large fund wanting to buy 500,000 shares without moving the price parcels the order across the day, roughly proportional to each interval's expected volume. The per-share average price they end up paying is their VWAP. Institutions then use that number to benchmark their own execution quality — "did my broker beat the day's VWAP?"

The chart overlay came later, as retail traders noticed that price respects the VWAP line like an unusually-sticky moving average. Same math, different purpose.

The formula

This part isn't in Kaufman — he treats VWAP as an algo, not a formula to plot. The standard definition everyone agrees on:

Where $\text{TP}_i$ is the typical price of bar $i$ and $V_i$ is its volume. For session VWAP, "anchor" is the session open — so VWAP resets at 9:30am each day and accumulates through the close. For anchored VWAP, you pick the anchor bar yourself.

The volume-weighted standard deviation gives the bands:

Plot $\text{VWAP} \pm 1\sigma$ and $\text{VWAP} \pm 2\sigma$ as envelope lines — Bollinger-ish, but centered on VWAP and volume-weighted.

Why intraday volume shape matters

Kaufman gives the key observation that makes intraday VWAP non-trivial:

In both stocks and futures, the greatest volume is near the open, the next highest volume near the close, and the lowest volume at midday.

The classic volume U-shape through the session means:

• The first 30 minutes of trading weigh disproportionately in VWAP
• Midday bars barely move it
• Power hour (last hour) pulls it back toward the late action

This is why intraday VWAP is meaningful as a reference — most of the session's volume (and therefore most of the dollar-weighted consensus price) is concentrated in the first and last chunks.

Play with it

Toggle Session VWAP vs Anchored VWAP, with or without σ bands. In anchored mode the slider lets you pick the anchor bar.

What to notice:

• Session VWAP (orange) tracks close to price through the day, with price oscillating above and below it. This is the line institutional execution desks are measured against.
• 1σ and 2σ bands widen when intraday volatility picks up. Price touching the 2σ band is rare — by design, since a volume-weighted 2σ is much tighter than a naive 2σ.
• In anchored mode, picking an anchor at a swing low means the AVWAP starts from there and becomes the "average cost for everyone who entered after the low." Until price breaks below AVWAP, the post-low buyers are collectively in profit. Break below — they're underwater. That's the behavioral significance.

Anchored VWAP — the modern retail extension

Anchored VWAP is not in our reference library — Kaufman, Murphy, and Bulkowski all predate its popularization in retail technical analysis. Brian Shannon is the most-cited modern advocate; his work elevated AVWAP from an institutional tool to a retail staple.

The anchoring logic is straightforward: rather than resetting at session open, start the cumulative sum at any event that mattered — earnings release, breakout bar, prior swing high/low, merger announcement, crash low.

The AVWAP from an anchor gives you the volume-weighted average price paid by everyone who participated since that anchor. If you're AVWAP-anchoring from the earnings release and price is above that line, the average earnings-reactive buyer is in profit. Below it, they're not. That's dramatically different information than a 50-day SMA — which weights every day equally regardless of participation.

Common anchor points in practice:

• Session open → classic VWAP (not "anchored" by convention, but mechanically the same)
• Weekly / monthly open → multi-day average consensus price
• Earnings release → what did the reaction-buyers actually pay?
• Prior swing low → breakeven line for everyone who bought the dip
• Breakout bar → breakeven for breakout participants
• News event / crash low → the "recovery" line

Core uses (editorial — not from our books)

Our reference library is silent on retail VWAP entries. The following are standard day-trading conventions, not citations:

1. Bias filter

Price above VWAP → session bias bullish. Below VWAP → bearish. Simple. Many intraday systems refuse to go long below the VWAP or short above it, period.

2. Mean-reversion entry

Price stretched to the +2σ or −2σ band often fades back toward VWAP, especially in range-bound sessions. A touch of the band + a rejection candle is a classic fade setup. Win rate varies by regime — don't use this in strong-trend days.

3. Pullback entry

In an up-trending session, dips back to VWAP often find buyers. Long the touch, stop below the next lower swing low. Works better than you'd guess, because institutional algos are literally engineered to buy at or below VWAP — they are the buyers at that level.

4. Execution quality check

If you trade manually and your average fill price is consistently worse than session VWAP, you're overpaying. Use it as a calibration metric on your own executions.

VWAP vs Moving Averages

Kaufman's one-line contrast via the TWAP comparison: "Equally spaced orders, which ignore liquidity, use TWAP, a time-weighted average price." VWAP doesn't ignore liquidity — it weights by it.

Rule of thumb: for intraday bias and execution, use VWAP. For multi-day trend, use an EMA or SMA. The two indicators answer different questions — they're not competing for the same job.

Limitations

• Intraday only by default. Session VWAP resets at the open. A "5-day VWAP" or "weekly VWAP" means different things on different platforms — some roll the accumulator, some reset weekly. Know your chart.
• Volume quality matters. VWAP inherits all of volume's problems. On thin stocks, one institutional order distorts the line. On halted or 15-minute-delayed data, VWAP is stale.
• Close-of-session skew. The final-hour volume spike can yank VWAP around late in the session. The 3:30pm VWAP reading doesn't necessarily reflect the day's prevailing average price — it reflects who was paying near the close.
• Session-boundary artifacts. On futures and crypto (24h or semi-24h markets), "session" is a convention — different brokers anchor at different times.
• Not a signal by itself. VWAP is a reference level, not a buy/sell trigger. Price touching it doesn't mean anything in isolation — pair with a price-action or volume-pressure read.

Hidden traps

• Treating VWAP as a moving average. Time-weighted and volume-weighted averages behave very differently on low-volume vs high-volume bars. VWAP barely moves at 11am (low volume); SMA moves every bar equally. These are different questions about different things.
• Anchoring VWAP arbitrarily. The anchor matters. AVWAP from a random bar in the middle of a consolidation tells you nothing. AVWAP from a meaningful event (earnings, swing extreme, news) tells you what those participants paid.
• Relying on VWAP for multi-day bias. Session VWAP resets. If you want multi-day context, use a moving average or a weekly AVWAP.
• Assuming the σ bands are normal-distribution percentiles. Like Bollinger (which we covered), the σ here comes from empirical data that isn't normally distributed. The "95% inside 2σ" interpretation is wrong.
• Ignoring thin instruments. VWAP on a 50k-daily-volume stock is meaningless if a single block trade moves the line.

Quick check

What you now know

• VWAP = Σ(TP × V) / Σ(V), where TP = (H + L + C) / 3. Cumulative since the anchor.
• Originated on institutional execution desks as an algorithmic order-placement tool and a benchmark for execution quality. Kaufman's contrast: TWAP ignores liquidity; VWAP weights by it.
• Intraday VWAP resets at session open; its shape follows Kaufman's observed volume U (heavy at open, heavy at close, thin midday).
• Anchored VWAP (modern retail extension, not in our books; Brian Shannon popularized it) starts the accumulation at any chosen event — earnings, breakout, swing extreme. Represents the average price paid by participants since that anchor.
• σ bands give a Bollinger-like envelope but volume-weighted and session-anchored.
• Primary uses: session bias filter, mean-reversion to VWAP, pullback-to-VWAP entries, execution quality calibration.
• VWAP ≠ SMA. Time-weighted and volume-weighted averages behave very differently. Use VWAP intraday, use moving averages for multi-day context.
• No statistical edge claim. Our reference books have no VWAP backtesting. Treat it as a reference level, not a signal generator.

Next: we close out Volume & Order Flow and open More Indicators — ADX & DMI is the natural start, continuing Wilder's 1978 toolkit from RSI and ATR.