Again,
Lets think about this from the valve point of view.
A cam is designed to give the valve a certain path.
Roller lifter, flat tappet lifter, mushroom lifter, no lifter...
It's all about the valve path and events.
If either roller or FT can provide the same valve path, there is no difference in performance. There is no performance advantage from one to the other. You're achieving the same goal, achieving the same performance. The lobe may be shaped differently, have different "numbers", but the end result is the same. Then, the choice becomes cost, maintenance, noise, or other non-performance attributes that may sway the decision. Up to the point where either can meet the goals of the build, I really don't see a way to compare them. They're different. Apple and an orange. Also, there is no benefit to opening the valve faster than needed. If the FT can do that (open the valve fast enough), the fact that a roller can have "more area" is meaningless. Once a FT can not open the valve as quickly and provide the area (duration) needed, then the roller becomes beneficial but if it's not designed right, a FT still might make more power. There are aggressive FT lobes that can make some serious power. Iv'e seen some that look like roller lobes, and there are roller lobes that are worthless. It's all about the valve path. The FT has limitations. That's just simple physics. The roller does too, but that boundary keeps getting pushed further and further with BIG lifter wheels and cam cores, but that's not what this discussion is about. At 4-500 hp (maybe more) and even up to 7000rpm, flip a coin. One is cheaper but has more risk. IMO that's about it.
Least that's how I see it.