Agreement over time

Pick targets and a reference set; each target's match-consensus percentage is plotted across assay history --record snapshots. Disagreement is literal — every precision, format-rendering, and error-code difference counts. (See compare for the present-tense, classified view.)

1 run on record. Lines need at least two points to draw — the chart will fill in as cron runs accumulate. Latest snapshot still classifies fine.

Per-run breakdown

Run	Target	Tests	Match	Some refs	Diverge	No data