# Addendum to PLAN_HAHN_CLOCK.md (/q C16, the floor), pushed BEFORE clock point 2
Delivered T2 moves between jobs at a fixed record age: kingston 0.910 -> 0.946 and marrakesh 0.907 -> 0.948 twenty minutes apart (referee:
about 4.5 and 3.5 sigma beyond shot noise); fez 1.051 -> 1.014 -> 1.000 over 0.7 to 1.8 h. So a single point's shift is not a trend.
Floor: the clock's slope is read against the spread of ratios between neighbouring points at near-equal age (the 0.7 h and 0.9 h fez points,
and the kingston and marrakesh run pairs). A slope is reported as BELOW 0 only if the decline from the earliest to the latest point exceeds
twice that between-job spread under BOTH fits; otherwise CONSISTENT with its bound, or CANNOT_TELL if the fits disagree.
Independently of any slope, each point is a delivered-vs-stated reading at a stated T2 age; a point at 8 h or later reading BELOW under both
fits is reported as such, with its age, under the bill rule of ADDENDUM_HAHN_POOL.md (fresh-record rule applies to the fez line).