Reason 1: With ID Norm off you can use the fact that Target ID increases with frequency as a tool to ferret out mixed ferrous and falsing ferrous targets. Specifically, on the Deus, steel crown bottlecaps tend to ring up like a high conductor in the high 80's low 90's (dime or quarter) depending on what frequency you are using. If the target is a true high conductor, if you increase or decrease the frequency (most easily done by setting up two custom programs adjacent to each other which are identical but differ in frequency such that you can easily switch between them using the +/- buttons) , a true high conductor's target ID will increase or decrease accordingly with frequency while a bottlecap or falsing ferrous will likely do the opposite of that when you change frequency. Nowadays, I mainly go by tone and tonal quality to bypass crown caps, but the ability to see how a target responds to different frequencies is a useful tool and allowing target ID to "float" with frequency can be useful. Obvious downside is that you have to mentally be aware that target ID will shift with frequency. That why I never memorize specific target IDs but ID "ranges" for common targets.
Reason 2: XP has chosen to normalize to 18 khz. This tends to jam all the high conductors target IDs up into the 90's and you lose the convenient wider spread of target IDs when using the lower frequencies.
Reason 3: The HF coils lack ID norm, so if you are switching between HF and LF coils, IMO it is better if they all behave the same with respect to target ID variation with Freq.
Caveat: The new X35 coils include 25khz which is above 18 khz, but ID Norm still normalizes to 18khz so using ID Norm only at 25 khz has the advantage of spreading the high end TIDs out a little (18khz reference) which gives more TID high end separation. Just an interesting little trick/quirk, but I still prefer to keep ID norm off across the board for all my custom programs and regardless of what coil I use. HTH