Here is why the OTS suits better than Registration in my situation:
Using Registration;
First Click (first step):
We have to click on the Regist section on the display (or by pressing the two Rgist Bank Select at the same time) to open up Registration Bank pages.
Second click (second step):
Searching/Finding and Selecting the Registration Bank through the pages.
Third click (third step):
Pressing/selecting the Registration memory button to reach the voice we want to play with.
Using OTS;
Only by one click [OTS 1-4] we have the voice selected.
That's why I prefer using OTS than Registration in my situation.
Interesting take, but I could also argue using OTS takes more clicks then:
1st click press the style select button
2nd search and find the style through many pages
3rd select if I want OTS 1-4.
As I stated; all user voices are just preset (or expansion) voices with a midi wrapper. OTS is giving you the "real" base sound, whereas Registration is hiding it because it's a snapshot of the system memory at that moment, including the filename of your modified sound.
Plus with registrations you can control the sequence of voices and styles, as well as tempos, key changes, etc. Much more power and flexibility (you can pedal through a registration sequence, keeping both hands free).
On the other hand, if you aren't as familiar /careful with Registrations, you could also accidentally switch output routing, vocal harmony settings, fine tuning, accidentally change tempos, accidentally change the mic levels, etc when you don't mean to. Using OTS and styles is more limiting, but also much safer (you can't actually save "global" values in ots. This includes the 'new new' for an edited user voice.
Mark