Anchored VWAP

Anchoring a VWAP to a meaningful event (a swing pivot, a session/period start, an earnings gap) gives a far more honest “fair price since the move began” than a rolling VWAP. Large, sustained deviations from an anchored VWAP tend to revert — which is exactly the Location evidence the reversal pipeline needs.

Features

• Anchor at an auto swing pivot (re-seeded from the real pivot bar, not the confirmation bar), a period start (session / week / month / year), or a manual timestamp
• Volume-weighted σ bands at two configurable multiples
• Distance-from-value in σ as a symmetric Location output (positive = above value → short location, negative = below → long location)
• Stretched / anchor / band context markers (Location flags, not entries)
• Light-theme dashboard and alerts

Anchor modes

Mode	Origin
Auto: Last Swing	most recent confirmed swing high or low
Auto: Swing High / Low	most recent confirmed swing of that type
Session / Week / Month / Year	start of the current period
Manual Date	a fixed timestamp input

Auto modes recompute the running sums from the actual pivot bar (pivR bars back) on confirmation, so value accrues from the real swing rather than from the lagging confirmation bar. The pivot only sets the anchor origin — it is not used as a trigger, grade gate, or dedup key (design skill §8).

Outputs

• AVWAP — the anchored volume-weighted average price.
• σ — volume-weighted standard deviation of the source around the AVWAP since the anchor.
• Distance (σ) — (close − AVWAP) / σ, the Location measure. |distance| ≥ Stretched at (σ) flags a stretched location.

These feed the Location role (skill §1) of a downstream reversal module as evidence, not as a blocking gate.

Notes

• Bands use the cumulative volume-weighted variance ΣpV·p / ΣV − VWAP² since the anchor.
• The VWAP line jumps at each re-anchor (session/pivot modes) — that is expected; each segment is an independent anchored series.
• In Manual Date mode nothing is drawn before the anchor timestamp.

Instrumente

VWAP is fundamentally a volume-weighted average price — there is no meaningful VWAP without real trade volume feeding the weighting.

• Valid: any symbol whose syminfo.volumetype reports base or quote — as a rule futures, exchange-listed stocks and crypto exchanges, plus broker feeds that pass real traded quantity through.
• Proxy: where that volume is a broker’s own share of the market (a CFD feed reporting base), the resulting levels sit where that broker’s volume traded, which need not be where the exchange’s did. Read them as approximate, not as the levels other participants watch.
• Invalid: any symbol reporting tick (a count of price updates) or n/a — there is nothing to weight.

On an invalid instrument the running VWAP/σ sums never accumulate real volume, so the AVWAP line, bands, anchor markers, stretched markers, and dashboard all go blank (na/—), and a one-time “benötigt echtes Handelsvolumen” warning label appears on the last bar showing the detected syminfo.volumetype.

Reference-market variant (documented, not implemented): unlike a pure oscillator, an anchored VWAP is an absolute price level. Computing it from a different market’s volume+price (e.g. NYMEX:NG1! while charting CAPITALCOM:NATURALGAS) would mix two feeds’ price levels onto one chart — exactly what DATA_VALIDITY.md §4.2 discourages: intraday reference-market price comparison ranks as “schwach” (Stufe 3) due to latency, session, and settlement mismatch between the two feeds. A dedicated single context module could carry a reference-market VWAP as its own series, but grafting a foreign price level onto the CFD chart inside this indicator is not sound. Not implemented here.