I've said this before but it depends how you look at it. As we know, information can be presented to reflect many different outcomes which is what makes statistical analysis so difficult.
Technically, if you include everyone in the population, and don't give any stipulations on how you're measuring tinnitus, then I could say without any doubt that tinnitus fades for nearly everyone. I can say this because most people get gig tinnitus that fades after a couple of days. Some can have it for up to a week before it fades, and for others it can be a bit longer. Obviously, this is being somewhat disingenuous and insulting to people who have a chronic case that hasn't changed in over two years or more.
So, when someone says it fades for most people, and then doesn't elaborate with a study or source, it can be freely interpreted as it gives no detail. Wouldn't it be better, if it were true, to say something like: for most people who have a chronic case of tinnitus (for two years or more) it's likely to fade over time. Otherwise we are free to add everyone into the equation and I personally know tinnitus fades because of the amount of times I've had it before it became chronic.
Clarity and distinction of exactly what is being discussed is paramount.