"Being in peril isn't fun. Pretending to be is." I kind of like that.
This is an excellent point, and one I want to expand on, because I think it has implications on two different levels in this discussion.
The first and more trivial level is that a tabletop RPG is a game, and while our characters may face mortal peril, the worst thing a player has to confront is the possibility of a papercut in terms of physical danger. So in the most obvious sense, a game is often about pretending to be in peril.
The second level, though, is the one the characters are on. I think most of us agree that a certain amount of in-character danger heightens the experience of the game by adding dramatic tension and fictional high-stakes action. We need antagonists, and those antagonists must threaten something we care about in order to motivate our characters to action. Let's call the stakes something like "little Suzie will be eaten by vampires." (And for the sake of the hypothetical, rescuing little Suzie is the point of this game.)
It's pretty clear that preventing little Suzie from being eaten by vampires is a good and noble goal, worth pursuing. But for there to be proper dramatic tension, we must also believe that interrupting the vampires' midnight snack will be a hazardous mission with perils aplenty--how else would our characters demonstrate their heroic attributes? If one PC can solve the issue with a phone call, or everyone knows that these vampires turn to dust when you merely brandish a crucifix at them, the outcome probably won't be very exciting. To be sure, you still want little Suzie to stay topped up on her blood supply, but there's not much drama if
no effort is required.
The key word in the above is "believe," however. The vampires in question could be paper villains that the PCs have a 95%+ chance of defeating (assuming reasonable competence on their part), but the action will be every bit as satisfying for those involved so long as the players' willing suspension of disbelief isn't violated. Therefore, it is beneficial for all involved if the DM talks up the potency of those vampires and the risk of failure for the PCs--and for the most part, the bigger disparity between perceived threat and actual threat the better, as long as the players continue to buy in to the perception.
If we want to maximize the perceived threat, and there's only so far we can talk up a differential between perception and in-game reality, why not just make the opposition very powerful, on a par with the PCs if not surpassing them? If the antagonists and the PCs are on an actual par with each other, exactly balanced in skill and competence, then the actual risk of failure is 50%. Nobody really wants to see little Suzie drained to a husk like an old Capri Sun juice packet, so making the actual chance of that happening 50/50 is rather harsh on both Suzie and the PCs, and a series of games where the overall PC failure rate was 50% is going to lose player interest quickly.
For that reason, I'd bet actual PC failure rate is closer to 5%, and in many groups, probably lower than 1% with stakes like little Suzie wishes she had. You only have to actually throw little Suzie to the wolves (different campaign: she's very unlucky, really) once in a great while to get a lot of perceived threat the next time she's in trouble--and it's even better if the failure can be used to motivate more competent effort on the part of the players next time.
TL;DR Part of the fun is pretending that make-believe characters are in more peril than they're actually in.
