The line drawing you made before painting that bust is actually very symptomatic, if surprising in someone of your technical skill. It's a tracing of contour.
It should have been a structural schematic, instead. You're recording lines and blots, and while focusing on lines and blots you are missing the whole point. Lines and blots are not the end, they are the means of constructing a representation of the form.
Man, you are trying to compete with the best, and annoyed that they are rejecting your efforts? Well, they are right. After seeing that trace I realized that your problem is far worse than I might have guessed from the finished work; your relatively polished technique threw me off. Now I can see it clearly. It doesn't matter how many drawings of this bust you made; if you aren't using the correct method, your effort is wasted.
Look at what you've traced. You've put in the irrelevant detail lines in the hair, where you should have been focusing on the masses it forms; and you've omitted relevant (and perfectly inferrable) lines in the helmet where they were obscured by the light bloom. You ignored the shadow despite it being useful for inferring the form. You weren't thinking, you were blindly following the photo and piling mistake upon mistake in the process, because you weren't looking at the right things.
Here, I made a quick example of what a proper starter drawing of the same thing might have looked like. Can you see the difference in the method? I am not following the naive lines, I am figuring out and reconstructing the forms. I couldn't care less if a particular line is visible in the photo; all I care is whether it helps me to record the 3-D form for later rendering. I put in center lines to record the angles. I use little horizontal marks to record the parallels. I put in symmetrical lines for every structure even if I cannot see them; if I see a particular gradient on the right, I use my knowledge of perspective and human anatomy to figure out what structure is making it, then I mark it with a form-describing line and recreate the symmetrical line on the left even if the camera did not register it. I do use the contour, but only where it describes the actual form, not where it is visible in the stupid photo. Above all, I am not trying to copy a photo; I am building a head by inference.
Is this enough of an explanation?