One of my great problems with running AD&D over the years has been the question of character power levels compared to challenges over a campaign scale. The basic problem is that any threat that is to be present throughout the campaign is either too great at the beginning or too feeble at the end. When planning a campaign therefore, you need to take considerable pains to work around this, typically by using an inverse Russian Doll, where the scale of the problem and the threat to the characters expands over the course of the campaign. This is frequently facilitated by having a heavy travel motif in the game: the characters travel from a small game environment to a larger one.
In less charitable or less well disguised environments this could be described as a shifting goal post. A shifting goal post is a fundamentally frustrating thing to interact with, so most games instead flatten the gradient of character improvement. However, I am certain that this is one of the great appeals of D&D: watching a minnow become a shark.
Where the idea of an expanding circle is for whatever reason inappropriate, the next most common solution is to more closely control the scale of the adventure to be contained within certain power brackets, usually in conjunction with re-grading the advancement profile of the characters. By “contained”, I really only mean that it doesn’t change a huge amount, not that it can’t be vast in scope. For example, an Exalted campaign might legitimately and comfortably encompass a wholesale reconstruction of the world as it’s campaign story, while still being described as “contained” in as much as the characters are within the same ballpark of powers at the start and finish. You might generally recognise the way in which super hero stories can be large in scope without needing changes to character power levels over the course of the story. A D&D example would be a campaign where the characters start at 1st or 2nd level and ends at 5th, or 13th level to 15th, and so on.
This is the easily identifiable part of the problem of scale, within the power of the GM to determine. The second part of the problem can be conveying the chosen scale to the players, especially in game environments that aren’t mechanically “balanced” by such factors as levels. A game like, for example, Mage: The Awakening is only very crudely adjusted for scale by considering “gnosis” or the total number of arcana a character has access to. The difference in appreciable game-effectiveness between a Mage with Gnosis of 2 and one with Gnosis of 3 is difficult to pinpoint. So too with the arcana. A well diversified character with 5 or 6 arcana at 2 or 3 dots is potentially insignificant compared to a mage that has only a handful of arcana at high levels.
The only “mainstream” game I’ve read that really tries to get to grips with the issue of scale, other than with the regimental D&D method of levels, is Unknown Armies, 2nd Edition. Although even there, I think the implication of it’s presentation is that you would start your game inside one of the three brackets and more or less stay there. I’ve never actually played it, so I’ll leave my speculations on it specifically there.
Synthesizing the above, and adding a dash of experience, I think that games need to settle the question of scale for three different areas. They need to decide what the change in power levels over the course of the game will be, what the range of player characters will be, and the scope of the story. Of the games that I’ve been involved in, I can think of only two where I was comfortable answering all three of those questions nearish the outset of the game, and was generally right. My more general experience has been that, like so many other aspects of the group dynamic, there has not been much shared clarity.
A couple of years ago, was trying to explain Exalted to me, and we touched on the issue of scale. The general regime we discussed was in this order: either the characters or the story should be in an epic scale, but not both. I think his preference at the time was also that small scale characters not exist in small-scale stories, but that seems a reasonably natural fit to me. Really broadly what this means is that if your character has awesome powers, the powers shouldn’t be important to their story, and if your character is an ordinary mortal they should be swept up in a story beyond their imagination.
I think we can see these ideas at work especially in the best super hero material, but the example I like to use is The Empire Strikes Back. I think the single most resonant line in the movie is “Luke, I am your father.” At stake for Luke in the movies is, yes, giant cosmic power; but more importantly, I think the commonplace choices of friendship and family are what really drive those stories. Many of those mechanisms are well used in daytime soaps, but the apparent powers-at-play in Star Wars distract us from their mundanity, and otherwise add an exotic flavour to them. The audience does not invest in Luke because he’s a bad-ass, but because we empathise with his ordinary human qualities.
In a previous post I said that “[y]ou should have in mind some kind of future for the character” you’re playing. I’d like to emend that statement to just say that when envisaging the future, you should bear in mind the relative scales of the story the character is in, and the game generally. Concomitantly, I think GMs should set a campaign scale and clearly discuss the limits of that scale with the players so that they know what the limits are.