One of the things that I've found frustrating about many games is the idea of Supernatural Powers within the context of the setting and the context of the existing game mechanisms.
Some games get the setting right, although this is most commonly the case when games are using supernatural powers adopted from a setting somewhere else. Perhaps seeking to simulate a particular run of comics, a successful movie or a series of popular novels. It's a no-brainer when you attempt to get something right, then simply copy another source verbatim...(don't ask me why Hollywood can't get it right when they try to cash in on the success of popular comics or award winning graphic novels...that's another tirade altogether.)
Getting the mechanisms right is another matter entirely. But the mechanisms depend on what the game is trying to achieve, and the achievements with the system are critically linked in with the agendas of the theme and style of play being sought.
This is purely my own opinion (based on what I've read over the years)...DC comics are about powers, Marvel Comics are about the flaws, many independent comics mix the two to varying degrees of success.
Translating this back into game terms, we need to consider how much of a game is going to be dedicated to the special abilities and cool tricks that the characters can perform. How much of the world can be transformed when these effects come into play?
Many games over the years have simply presented a series of powers, ranging from a dozen or so, through to the hundreds of powers available in a game like Heroes Unlimited and its numerous accompanying sourcebooks. This method isn't so much a mechanism as it is a smorgasboard of choices, some of which are far more lucrative than others (simply inviting certain competitive types of player to mass produce a series of clones with the same ultimate combination of game manipulating effects).
Cetainly not what I'm after. It doesn't engender creativity, and it doesn't maximise the potential enjoyment for the widest number of people. In a competitive atmosphere, all players end up with vague copies of one another; while in a narrative environment, one player gets the limelight while others are often relegated to sidekicks. That would be looking at things from the power perspective.
Looking at the flaws and weaknesses definitely brings more humanity to the story, and personally I think that's a far more interesting way to focus a game. Look at what the powers can't do, rather than looking at what they can. A good GM will design their game to ensure certain players can't always succeed and will give the less obvious characters a moment to shine.
But how do we develop a game mechanism that brings this about?
For this I propose a system where characters need to engage in their flaws before they can access their powers in a way that is meaningful for the progression of the game. (Note that I chose the grammatical construction of that sentence carefully).
I haven't said that players only get access to their character's powers if they use their flaws...a character with super strength will always have super strength, but they won't always find the right opportunity to use it, especially not when the situation calls for subtle diplomacy.
In certain styles of game, it might be hugely advantageous to possess mind control compared to possessing an ability to turn into a gerbil, but if you watch the movies "Sky High" or "Mystery Men", you'll see how under-rated powers can really be used to save the day if they are applied at the right time and place.
Consider every other medium where supernatural powers are used and you'll find that the better examples show the abilities in a context. Everyone takes for granted the fact that Superman can fly, has amazing strength and has an array of powers that would just seem cheesy if he were produced today...it's the times when those powers are of little use to him that the memorable storyline develop (or the times when he has to combine various powers and push them to new limits). A vampire may be known for its suave and sophisticated attitude, its ability to read minds and manipulate the emotions of its victims, but it's the vulnerability to sunlight that defines the race.
It sorts of brings me back to an earlier mechanism, where experience is only gained through failure. But I'm looking at it from a different light here; in this exploration I'm more interested in how players can advance the story through the strengths and weaknesses of their characters, rather than considering how the players can advanced their characters through the successes and failures within the story. Almost like two sides of the same coin, but not quite.
You can probably infer from my comments so far, that I'm not going to generate a list of powers for this game mechanism, I'm going to keep the specific powers fairly abstract and focus on how the powers could manipulate the world within the game.
For this I like games where degrees of success are measured; one success being a minor success or a partial success, two being a full success, three or more adding degrees of "awesomeness" to the final result. On the negative side, there needs to be at least two degrees of failure; one where the character has simply failed to make significant headway in their task, and another where the character fails catastrophically.
There are plenty of systems in the gaming comminuty where this scale of results can occur.
We also need some kind of mechanism where the scenes are generated at least partially on-the-fly, because we can never tell how characters will use their powers or what weaknesses might become significant.
Basically, I'd look at a series of counters to reflect a karmic balance. Every time a player accepts a penalty, they get the opportunity of unleashing one of the strengths later on. But there would have to be some kind of cap on this.
Let's look at this from a communal story situation. Scene is vaguely set by the GM, any of the players may introduce into the scene some kind of issue that has to be resolved in addition to the premise currently laid out. A player may introduce situations that make it harder for their own character, In exchange, their degree of success is reduced by one level (marginal successes become marginal failures, and marginal failures become catastrophic failures). For confronting an obstacle, the player now gets a token that may be used to fuel their character's power later on. When they later use their power, they may narrate a part of the scene to show how their power will be of use (The GM narrates that a crash victim needs to be rescued...the player spends their point then describes that a half-crushed car pins the leg of the occupant, the character's super strength can be used to rip the vehicle apart and release the victim).
As long as the GM is fairly loose with their needs from the scene, they can keep the narrative open and can allow new elements to be added by the players.
If we look at it from a competitive perspective, then players might be able to add elements to the scene that will hinder one another. To keep things simple, the same meta-game currency would be used. A player can use one of their earned tokens to apply a bonus to themselves, or can introduce an element into the scene that will hinder one of their fellow players. In this set-up, a central pool of tokens is absorbed by the players as they write their own weaknesses into the storyline, and if players declare situations where their strength would be an advantage then they put the tokens back in the central pool...but the twist is that a player can give a token to someone else. Effectively adding it to the central pool, then forcing another player to take it by introducing their weakness. The combo move is less effective than the combined sum of transferring the token to bring a power into play, then transferring the token in exchange for a weakness, but it makes the players more wary of one another. This element of paranoia can really give a game a darker vibe.
The mechanism could even be combined with other mechanisms suggested so far to give some really thought provoking concepts.
Combined with the first mechanism, the conscience bag, players could be forced to consider whether they really want to use their powers or not. Draw a white bead and you decide that the use of powers in this situation would be extravagant and you would be better off saving them as an ace up your sleeve later in the story, draw a black bead and you say "damn the consequences" and go all out with whatever you have at your disposal. Perhaps if you draw black beads from the bag, you'd have to use as many of your accumulated power tokens as the number of beads drawn. The darker your urges, the more incline you are to use the power to its fullest extent.
That helps to combine a suitable capping option with the mechanism, but having more than one type of token floating around can confuse players who aren't used to the concept of meta-game currencies...trust me, I've now run enough games of The 8th Sea to see the confusion on people's faces.
If degrees of success are added by the judicious application of powers to a suitable situation, then another method for capping the degree of power could simply be a degree of mastery in the power being used. One player might have a clairvoyance of 2, another might have a clairvoyance of 7. The first player can only spend two of their accumulated power tokens on a task where their clairvoyance skill is useful, while the other player might be able to spend seven tokens on the same feat. Presumably the character with only 2 levels, would have a variety of other powers through which their tokens could be used, and this would be tied in with the experience systems of the game.
I guess the idea presented here could easily be used to add unconventional powers into any type of game, giving them a different feel to any existing abilities. Maybe psychic powers into a modern day spy genre, or freeform magic into a structured world of super-heroics.