Just because something is theoretically possible, doesn't mean it's doable for human beings. So in the end, even with frame data you'll have to feel it out and see whether or not
you can do whatever trick you had in mind.
I really don't see the point of it all and unless you have a really good example where frame data helped you out tremendously, I don't think I ever will. And if you don't, well then it kind of proves my point.
well just consider the up-b edge stall - frame data tells us it's 100% invincible when done 'well' (it doesn't even need to be done perfectly), whereas that's a very difficult thing to know via practice (because even if you're just 90% invincible, it's possible that people just keep whiffing while you're practicing, or you chalk up getting hit to you making a mistake)
and knowing which characters can or can't jump out of fox's uthrow uair or falcon's throw->knee is HUGE
i feel like i gave examples in my first post so i don't know why you had to prompt me for more -_-
the gist of it is that knowing that X move has 10 frames of lag (1/6 of a second) is not important in itself - nobody's just counting frames and thinking that that specific knowledge is important
knowing that X move has 10 frames of lag, while your opponents' potential punishment comes out in 8 frames instead of 12 frames IS important... frame knowledge isn't used in a vacuum; it's used to compare situations (i.e. if you see the thread in the falco boards about the frame advantage/disadvantage in a dair-shine-dair pillar, you can see the best times to do your dair, how mixing it up exposes different vulnerabilities, how you're 100% grab-proof vs. zelda, etc)
frames are just a concrete way to determine "more" or "less" laggy/effective/whatever