[solved] Why so many scores with poorly adjusted sound levels?

I wonder why do we have so many scores uploaded to musescore.com that have poorly adjusted sound levels?

Take for example this piece:

https://musescore.com/user/1889291/scores/4823813

My immediate impression, when I had first listened to it, was that some instruments in it were way too loud at some spots. Then I suggested that that was exactly what the author had intended.

But then I noticed that he also provided there a link to an audio file of the same piece:

https://soundcloud.com/verywosushi/euphonium-variation-waltz-guy-bacos

When I listened to that file I realized that my original guess was absolutely correct – all the instruments in that audio file sound much softer then on the score playback on Musescore.

And that was not the first time when I had bumped into such cases.

I have some of my personal guesses here:

1) No matter how diligently you fine tune the instruments on the mixer while creating the score on your Musescore on your computer, all the set levels will be gone after you have uploaded your score to musescore.com ;

2) No matter how diligently you set up the dynamics notations on your score, they will still not be played properly on musescore.com (for some technical reasons, for example, the version change). So there is no need to bother;

3) Usually, it is solely at the director's discretion what amount of loudness to give to this or that instrument, so authors just trust the taste of any director who will hopefully one day pick up that author's score to play;

4) (similar to point 3) While meticulously specified dynamics will work well for the playback on Musescore (on your PC or on the website), that will look a bit ridiculous on paper and, perhaps, even frightening for a living conductor who is not a score playing-back machine. So it is more practical just to keep the written score as "clean" and "conductor-friendly" as possible;

5) (similar to point 4) It is a good practice, besides sharing the score, to also provide a link to an audio file showing exactly how the author wants that score to be played (like in the example above), in which case there is again not much of a need to bother properly specifying dynamics on the score;

6) None of the above? Any other possible reason then?