01 / Methodology
DQ / METHODEvidence before
conclusion.
Market Intelligence
Infrastructure
Infrastructure
DALQEN preserves source identity, freshness, coverage, and disagreement as market facts become structured intelligence.
02 / Evidence Model
A documented path
from fact to context.
The system may conclude that evidence is stale, mixed, unavailable, or insufficient. Those are valid outcomes.
- Source facts and derived intelligence
- A source fact is retained with its provider, timestamp, unit, and coverage. DALQEN is identified as the source only for analysis or research produced by DALQEN itself.
- Freshness
- Observation time and ingestion time remain separate. A valid daily or scheduled observation is not described as real-time, and an old but reachable source is not automatically described as current.
- Evidence confidence
- Confidence considers freshness, coverage, data quality, independent support, and agreement or disagreement between evidence.
- Duplicate protection
- Stable provider identities, timestamps, and methodology keys prevent the same observation from being appended more than once.
- Stale-data handling
- A failed or late check does not create a replacement value. The last valid observation keeps its original timestamp and may be marked stale.
- Conflicting evidence
- DALQEN preserves disagreement. Mixed, inconclusive, or insufficient-evidence states are used when the facts do not support one conclusion.
- Methodology versioning
- Changes to a calculation or classification receive a new version. Historical records keep the version that produced them.
- Unavailable data
- Coverage may be incomplete and source timing may differ. Missing or stale evidence degrades an assessment rather than being replaced by a fabricated conclusion.
- AI-assisted analysis
- AI may assist classification or contextual analysis only when explicitly enabled. Its output must remain distinguishable from verified source facts and is not presented as autonomous market prediction.
01
02
03
04
05
06
07
08
09
03 / Confidence
Confidence reflects evidence quality, not return probability or win rate.