Fixed fractional spectral width gives no clear error gain in 103 selected CHIME bursts

We compare two spectral-width rules for the first and last fitted components of 103 selected CHIME/FRB Catalog 2 events from 47 sources. Fixed fractional width gives a 9.32 percent error reduction against fixed width in MHz. The 95 percent source bootstrap interval is -9.59 to +28.27 percent. A 20 MHz band margin changes the gain to -3.49 percent. Exploratory repeater and one-off differences also weaken with that margin. Thus, these data give no clear prediction gain. The test uses measured endpoint frequencies and fitted spectral shapes. Beam response, band selection and shared fit errors prevent a physical conclusion. No new source or planet is found. The bundle contains original code, aggregate results, a figure, exact input hashes and retrieval steps. External source records and event-level derivatives remain outside the bundle.

astronomy · CHIME · fast radio bursts · spectral bandwidth · fractional width · Doppler scaling · null result