Which providers feed each stat, how reliable each has proven, and what happens when they disagree.
Knox blends several data providers. Each one earns a 0-100 reliability grade from its live track record: how often it agreed with the cross-source consensus versus how often it was the rejected outlier. Identities are redacted — the grade, not the name, is what matters. A grade is shown but not acted on until a source has enough observations to have earned it.
src_e4b7a8a7
Supplemental · won 100% · outlier 0% · 49 obs
src_8e95c413
Supplemental · won 0% · outlier 100% · 49 obs
The oracle already down-weights unreliable sources inside its Huber-robust triangulation; this page publishes the grade that judgment produced. The same grade is bound into answer receipts so any answer can be checked for which sources it leaned on. Conservative by design: a low grade can cost a source weight, never inflate it past parity.
Technical term: oracle source honesty
Knox blends several data providers and weights them by how reliable they have proven. This page publishes both halves of that judgment: whose word Knox has come to trust, and every value it rejected as an outlier. It is powered by the live oracle disagreement log — real movement, not the replayed receipts our corpus-health check flags.
See the source reliability scorecard → graded 0-100 →Trust weights frozen — 1 distinct disagreement replayed across 48 rows (98% replays). Weight drift reflects repetition, not new evidence.
The drift below is real arithmetic over the log, but it accumulated from 1 distinct disagreement replayed across 48 rows — repetition, not 48 independent confirmations.
Fields where data sources have clashed most often (recency-weighted) — the stats most likely to disagree next, so Knox triangulates a third source there before answering.
Knox has rejected src_8e95c413 on points for 161 consecutive disagreements across 91 days.
Honest caveat: those rows are the same disagreement replayed, so this is one standing call — not 161 independent confirmations. It will become a live streak once the disagreement log varies.
Fields where Knox's data sources disagree most — ranked by dispute magnitude. Based on 161 oracle observations.
161 disputes · 1 sources involved · avg deviation 6.34
One source drives 100% of disputes on this field — Knox triangulates around it each call.
Dispute magnitude = disputes × (1 + mean deviation from consensus). Higher = Knox works harder to triangulate. Provider IDs are redacted on this public view. View full outlier feed →