An early width fit reduces later tail-time error by 59 percent for 35 selected CHIME bursts of FRB 20220912A

We test whether pulse shape predicts fitted scattering tails in CHIME/FRB Catalog 2 release 20251231. A log-width relation fitted to 39 early bursts reduces mean squared log-tail error by 58.6 percent on 35 later bursts. Each part spans 30 observing days. Separate resampling of early and later days gives a 95 percent gain interval of 38.1 to 72.4 percent. Single-component and high-signal cuts retain the gain. A time-only fit increases later error by 21.0 percent. The result shows a repeatable relation between fitted parameters in this selected sample. Width and tail share the same pulse fit. Selection, fit errors, intrinsic structure, and propagation can cause the relation. It does not confirm a change in plasma or a new source. Independent FAST work detects a scattering screen for this source but does not confirm these CHIME tails. Code, frozen source rows, input hashes, figures and offline steps are included.

astronomy · CHIME · FRB 20220912A · scattering · pulse width · time validation · fit covariance