Not in my experience, unless you use “Loud” setting.
Actually, I captured the output of the app when playing “Anne Sophie-Mutter / Vivaldi: Le quattro stagioni / 12. L’inverno: III Allegro”, which has:
Loudness, Sample peak, RMS
-19.0 LUFS, -0.49 dBFS, -24.24 dBFS - capture_norm_off.flac
and I got literally the same data, to the sample, with normalization turned on.
Sorry for a car analogy, but that’s like saying that cruise control makes zero sense as long as you have a functioning gas pedal.