I am sorry but you are completely incorrect here. If you publish to Arxiv you can't publish in a journal that actually gets you the "points" you need to advanc

I am sorry but you are completely incorrect here. If you publish to Arxiv you can't publish in a journal that actually gets you the "points" you need to advance your career or secure your funding. The only exceptions are domains like mathematics or theoretical physics where publishing pre-prints and sometimes full papers on Arxiv is the norm. Most research fields are not like that. One of the first things most journals will ask you to do is an exclusive copyright assignment as a condition of publication. The only thing you can do afterwards is to publish the paper on your own website (and often not even that, at least not the version that the journal has a copyright on) and use it privately e.g. for teaching or further research. Nothing more. So there is no way to publish the same text both in e.g. Elsevier's journal and on Arxiv without risking getting sued. And few people have the time (unpaid, of course) to write a second version of the same paper only for the sake of uploading it to Arxiv. The part about the rating systems - this hasn't been created neither by scientists (who work mostly on reputation) nor publishers. This pressure to score, evaluate, grade research output is a purely governmental thing - bureaucrats and politicians wanting a simple number they can compare and use to decide on who gets and who doesn't get funded, who can and who can't apply for tenure and so on. Elsevier and others (Scopus, for ex.) only filled the demand by providing easy to use statistics, such as impact factors, citation numbers and so on based on their databases. Oh and they have very good lobbyists and e.g. in Germany Elsevier was put in charge of all this data tracking for the German government ... However, none of this requires any of the publishers - all publications are recorded by the universities anyway, citation numbers are quite easy to estimate even by googling (Google Scholar does that, without any access to the proprietary databases). And getting stuff like impact factors (which are a largely arbitrary metrics comparing cited vs all published articles in a certain journal) is in the interest of publishers because otherwise their journals will be irrelevant.