Guides, manuals and platform references.
Yesterday's VWAP can finish far from today's opening line without anything being broken. A session reset discards the previous session's accumulated inputs. The new line answers a new question: where is the volume-weighted average since this session began?
That distinction matters when a trading rule compares price with VWAP near the opening boundary. You may be comparing two different measurement windows while treating them as one continuous reference.
VWAP divides accumulated price times volume by accumulated volume. TradingView's standard indicator defaults to hlc3, the average of a bar's high, low and close. Its Anchor Period controls when those totals restart. Session, Week and Month therefore describe different accumulation windows. See the official VWAP settings and calculation.
Here is an invented three-bar example using an arbitrary price unit. It illustrates arithmetic, not market observations or a backtest.
| Bar | Price input | Volume | Price × volume |
|---|---|---|---|
| 1 | 100 | 10 | 1,000 |
| 2 | 102 | 30 | 3,060 |
| 3 | 101 | 10 | 1,010 |
The total is 5,070 divided by 50, giving VWAP of 101.4. The larger second bar has more influence than either smaller bar. Taking a simple average of the three prices would answer a different question.
Suppose the next session's first bar has a price input of 104 and volume of 5. After a reset, VWAP is 520 / 5 = 104. If the anchor instead retained the earlier bars, the calculation would be 5,590 / 55, or about 101.64. Neither number is a correction of the other. Their starting points differ.
A session anchor suits a question about the current trading day; a weekly anchor suits a question about the week so far. Changing the anchor changes the question. It should not be a last-minute way to make a disappointing signal look agreeable.
TradingView advises using an anchor period containing several chart bars. Session VWAP on daily bars resets every bar and loses the intended intraday comparison. Also record the Source and calculation Timeframe when comparing two charts; matching the indicator name alone is insufficient. These controls are described in the same VWAP reference.
A manually anchored VWAP is another choice: TradingView's Anchored VWAP drawing begins at a point selected by the user. A line anchored to an event is not expected to behave like a line restarting each session. Label the anchor in your notes so that a later review does not silently compare different definitions.
Do not diagnose a missing reset solely by the clock passing midnight. TradingView documents that overnight trading days can cross calendar dates, and US futures holiday schedules can combine shortened sessions into one extended trading day. Session-based indicators may consequently retain their accumulation across an intervening session break. This is documented behavior, not automatically an indicator fault. Check the instrument's daily bars and the relevant exchange schedule. See TradingView's explanation of extended daily sessions.
If you write your own session logic, define its clock separately. The Pine session timezone and daylight-saving guide explains how to make that rule explicit. A custom clock boundary and an exchange-defined daily session should not be assumed to mean the same thing.
Consider a hypothetical rule that buys whenever the close crosses above VWAP. On the last old-session bar, it compares the close with the old accumulated average. On the first new-session bar, it compares the close with a freshly started average. A cross spanning that boundary mixes two reference windows.
Decide in advance whether that boundary comparison is allowed. One testable alternative is to require both comparison bars to belong to the same anchor period. Another is to wait a specified number of completed bars after each reset. Neither rule is inherently more profitable; each expresses a different strategy and needs its own evaluation.
Before testing, write down the exact symbol and feed, chart session selection, chart timeframe, VWAP source, calculation timeframe, anchor, and first eligible signal bar. Then inspect several session boundaries, including any holiday period in the sample. For each suspected cross, record the two closes and the two VWAP values. Those four numbers are more useful than a screenshot caption saying that the line “jumped.”
A reset starts a calculation. It does not, by itself, supply a reason to trade.