It's been said, many a time, that artists are their own worst critics.
We know what we meant to put on the canvas (page), the effect it's supposed to have. Did it work? Did it not?
We're said to be too close to our work to know.
This is why writers rely on beta readers and/or critique groups. Rip it apart, spit it out, say what works, say what doesn't.
I've heard it said that writers can't evaluate their own work, but I'm unconvinced. Often, I've noticed that when I'm unsure about something, betas really like it.
But when I'm 99% certain something doesn't work? Betas agree with me.
And those writers I edit? If someone says "I think there's something wrong with the structure," well, there's something wrong (or at least unusual) about the structure. (I don't go into a work expecting to find those problems, either—I actually forget the author said anything until I start noticing the problem. Then I think, "Hope the author's okay with hearing there's structure problem—oh, right. The writer said she was worried about that.")
Now, as a writer, when I'm nervous or unsure about something, I've found it's usually best to leave it alone.
But when there's certainty? Fix it.
Sure, there are exceptions. It also differs from author to author, and likely even from story to story.
But with practice—and that means practice evaluating new stories, not the selfsame one over and over—you can get better at identifying when there's a specific problem.
Then you hand it off to a trusted beta reader (or however many you have) to see if the story did what you wanted.
If it failed, you can try again, or you can accept the story for what it is.
If it succeeded, you can rejoice and hope the next one will do the same.
But if you find yourself completely rewriting every little word and phrase because you're certain it isn't good enough, stop. There's something called over-editing, and the results ain't pretty.
Note that I am not saying an author can accurately evaluate the quality of something s/he writes. I'm saying an author might be able to accurately evaluate the major problems in a piece s/he writes.
We can't judge our own quality, because we either see or overlook every flaw, real and imagined.
But maybe we can judge when our intended comedy story ended up with a despicable heroine.
Do you find that your self-evaluation of story errors matches what your beta tells you? Or, if you're a beta reader, do you find that the writer(s) you beta read can accurately judge the errors in their own work.