Thanks for your comment, Stephan! And I actually agree with you. Maybe my point wasn’t very clear (I feel disturbance in the force :D). Let me clarify. I’ve faced many cases, when a reviewer who was supposed to be an expert in the area (because they accepted to review the paper) did not know much about forecasting and did not even make an effort to understand the paper and judged it hastily, providing comments like “you do not cite papers on forecasting with high frequency data, so I recommend rejection” (this is one of the comments I received for CES). So, this is just a sloppy revision, and my point in the post is that the reviewers need to understand that revision is a serious job. Yes, a reviewer should look critically at the paper and they should help making it more understandable if it is not well written. But the revision is a process done by two sides, not just one. So, if the paper gets to the fourth round without progress, this means that the reviewer and the authors are speaking different languages, and both sides should make an effort to understand each other (preferably, much earlier than on the fourth round). However, there are reviewers that do not want to make that effort and prefer just to get rid of the paper, so that they do not need to do the job, but at the same time can claim that they review papers in this and that journal.
And yes, there are good reviewers as well, I had several during this journey. They were responsible and helpful. And yes, the authors should write papers which are easy to understand. The points above mainly apply to those reviewers who think that doing a sloppy job is fine.

I disagree with one point: if I as a reviewer do not understand a paper, it is usually not my responsibility to invest a lot of time. After all, I am supposedly an expert in the field. (If it turns out that I am assigned a paper in which I am not an expert, then I should notify the editor and withdraw.) Thus, I am precisely the target audience of the papers I review. And therefore, the *author* needs to invest every reasonable effort to make their paper understandable – after all, if the reviewer, an expert, does not understand it, how will later readers understand it?

And yes, there is frequently a tension between “I do not understand X, but the paper is good, so please explain X better” and “I do not understand X, and the paper is generally weak, so I recommend rejection”. I have heard it said that the job of the reviewer is to weed out bad papers and make good papers better. One can err in either direction. But nobody is happy if I get a paper I believe is a publishable-if-better-explained one, and we then have three review rounds before the editor and I are convinced that the author is *not* capable of explaining their idea better, and then the paper is rejected after three rounds.