Question about general practice

Hi,

In the lecture, the preferred method is scaling before fitting the data. After identifying the worst-performing features, we omit them and run the fitting again.

Another question: are there situations where scaling does work? Do we do scaling all the time?

Thanks