Singapore Gold Market Intelligence

Gold preserves immutable observations of the Singapore gold market through time, reference prices, dealer prices, premiums, spreads, storage fees and provider terms. This is a historical record of how the market has actually behaved, not a forecast.

Data sources:Local browser activityObservations are immutable records with full lineage. Only validated observations appear in public histories.

Observations

Immutable records

Market changes

Genuine source-observed

Corrections

Normalization/parser

Observation gaps

Cadence breaks detected

Monitoring coverage Local browser activity

Dealers monitored

Products monitored

Valid observations

Change events

Gaps detected

No dealer observations have been collected yet. Observations accumulate as the scheduled adapter pipeline runs. Gold never backfills missing observations, monitoring begins when collection starts.

Adapter health Local browser activity

Adapter	State	Last success
LPM Singapore	healthy	—
BullionStar	healthy	—
GoldSilver Central	healthy	—

No observations have been recorded yet in this session. Observations are collected as the Gold reference price feed and dealer product data are accessed across the site. This is a persistent intelligence layer, over time it builds a genuine historical record of how the Singapore gold market has behaved.

Gold never backfills missing observations. Every record carries a source, timestamp, verification state and lineage.

How this intelligence works

Immutable records: Observations are never mutated in place. When a value changes, a new observation is recorded and the old one is marked "superseded".
Source vs derived: Source-observed values come directly from feeds. Derived values (premiums, spreads) retain the IDs of the observations they were calculated from.
Lineage: Every chart point or Goldie claim can be traced back through source → normalization → reference → derived metric.
No backfill: Gold never fabricates missing observations. Gaps are flagged, not filled.
Change semantics: Genuine market changes are separated from normalization/parser corrections.
No inferred timing: The exact time a value changed is not inferred from intermittent observations. effective_at is only set when the source explicitly establishes it.
Adapter failure ≠ market event: A failed adapter means Gold does not currently have a fresh observation — not that a product was removed or a price became unavailable.
Price tiers preserved: Quantity and payment-method tiers are never collapsed into a single price.