Earlier quoted context omitted.
I left my PhD in computational geometry for similar reasons. I was told flat out by my advisor that the chances of me getting a paper accepted in even a middling journal were low. Not normally low (as in tough review), but so low because the entire field is dominated by a handful of very high profile researchers and mathematicians that essentially gate keep knowledge on it. It was also necessary to publish some paper…
One slight saving grace (after having graduated) is to feel fine writing documents like papers and just putting them up on https://arxiv.org/ under my own terms. One additional technique I learned from others is to use submissions to conferences as a way to get feedback on your work without any expectation that it gets accepted. You do that early knowing there are flaws, but do it to get early warnings from reviewers…
This is, of course, one if the problems with current research as it wastes everyone's (yours, reviewers, ACs, etc) time. In particular, because you're not guaranteed practicable or actionable feedback and because you're increasing the time reviewers must review work (i.e., more papers to review as submissions are higher and quality lower).