I think you misunderstand. I did not say that there should be no losers. More importantly, whether they should have losers is besides the point as it is a simple empirical observation that the Thompson Reuters IF goes up over the years.
No, not for all journals, and even for the successful ones not uniformly every year. But the decent mid to high level journals in my field all have higher IF now than five years ago, probably simply because people cite each other more often. (Plus perhaps a bit of gaming the system by authors and editors alike.)
