Source Mesh
Source coverage
One instrument → many observers → one canonical market state. Market Atlas does not trust a quote: it observes the market through independent paths, counts source families (not URLs), and publishes the consensus with the evidence. This page measures that redundancy and lists where it is still thin.
Source redundancy
Source coverage →33of 235 quoted instruments observed through ≥ 2 independent source families
10 ≥ 5 sources15 ≥ 38 2197 single source
Weighted redundancy score64.0%
- Tier A majors covered
- 5 / 28
- Tier B liquid covered
- 27 / 128
- Proxies confirming
- 6
- Validators confirming
- 0
18% · target ≥ 3 families
21% · target ≥ 2 families
stablecoin markets, crosses
restricted-rights cross-checks
Redundancy targets by tier
By asset class
quoted instruments · share observed by ≥ 2 families · mean independent familiesSource expansion queue
0instruments below their redundancy target, most valuable gaps first (tier weight × missing families)Every instrument in this selection meets its redundancy target.
A family is a group of sources believed to share an upstream (declared, or inferred statistically). Proxies (stablecoin markets, crosses) and validators (restricted-rights sources) confirm a value but never count as independent families. Score = 65 % family attainment + 35 % observation attainment.
Likely shared upstreams
pairs of sources whose values coincide within 2 s on ≥ 92 % of ≥ 120 aligned samples — counted as one familyNo pair of sources currently looks like a shared upstream. The lineage engine recomputes every five minutes from live observations.
Quoted instruments
235
≥ 3 families
25
Tier-A target
≥ 5 families
10
most observed
No fresh source
5
last known value shown as stale