I clearly cannot write editorials (see here, here, etc.)

Can anyone please actually explain me how to write editorials properly? What I do now is just providing my thinking process to reach the solution, but it seems that most people have a very different thinking process.

UPD: after reading the comments, I think I got what's wrong in my editorials (and I tried to fix that, especially in the editorial of 2018B - Speedbreaker). One of the main issues is that I tend to skip algebraic manipulations, inequalities, etc. (i.e., steps that require simple algebra and/or mathematical logic), and it turned out that they can be the bottleneck for most people here. Lesson learned: if some numbers / variables appears from nowhere, I should try to elaborate, even if it's a single mathematical step.

For earlier problems, maybe I should also focus on how to convert the ideas into code.

Anyway, I think this comment helped me a lot to write a "better" editorial. In an ideal world, I would ask testers to read the initial version of the editorial and write similar comments, to help me "fix" the editorial. But during problem preparation, the editorial is the most neglected part for obvious reasons (there are a lot of little things that can ruin a contest, but in theory a contest can be held without even publishing an editorial), and ideally I don't want to fix the editorial 2 days after the contest end, because most people would have already read it. A solution might be just to give write access to the editorial directly to the testers and let them fix it if they had trouble understanding something?