If I were creating a character for a classic dungeon crawl where the whole point is to fight monsters for fun and profit, then I shouldn’t make a Barbarian whose fear of her own rage convinced her to be a pacifist. Now this isn’t to say that you couldn’t make a pacifist Barbarian, as that could be a fantastic character concept in a different kind of campaign (even certain kinds of dungeon delves) only that it won’t work in this particular campaign unless we toss the essential premise (kick in doors and hit stuff) out the window in favour of something else.
So what can a Game Master do to ensure that everyone knows what the campaign is about so that they can take that into consideration when creating their character?
Firstly consider whether what you intend to make is something they would be interested in playing. You should look to create something different for a party of combat heavy munchkins than you would if everyone was a budding investigative sleuth. This isn’t to say that you can’t include investigation among the combat or combat among the investigation (variety is the spice of life), but that a game that punishes combat or blocks off investigative routes just won’t be as satisfying for them. They’ll keep trying to creep off the edges of the social contract by instinct alone or will sit around bored and neither option is very fun for anyone.
Once you know that the game in question is something they might be interested in, begin with a chat about what the campaign is going to be about. While you can talk a bit about style and setting, the most important and so often forgotten element is how the game is best played.
Is this a conspiracy game fraught with hidden peril where even the other player characters might be out to get you? This sets out a very different kind of social contract than a game of silly goof balls where people jokes around and use their super powers in weird and wacky ways.
Then talk about the sort of tactics which could be particularly useful in this kind of game. While rule books give people a vague idea, they’ll only take you part of the way. Will the conspiracy be primarily solved through clue hunting, resource gathering, social maneuvering or violence?
The point of this talk about tactics isn’t to create a comprehensive list of what the players can do (hello rail roading!) but to give them an idea of what could work. And definitely get them to ask questions and pitch ideas. Is creeping around through air vents a viable option? How about nonlethal tactics? Would they work? Would not killing the bad guys only feed into the villain’s plans or will it help the PC’s bypass obstacles while keeping their integrity intact? Is the police department well-funded or corrupt? Be willing to compromise.
Maybe your game world has a well-funded police department and skilled ballistics team when a murder happens but they just don’t pay all that much attention to high speed chases. Now be aware the point of this exercise isn’t a series of Thou Shalt Nots. The point is to ensure that everyone knows what consequences this game is running with and the easiest points of contact. If the players know that an assassination will be seriously investigated, they will know to plan out their murders with greater care and perform them more rarely.
It doesn’t remove creativity to set up a campaign in a particular style and setting for players who could enjoy that. It enhances it. Get them onboard, talk about your various game expectations and then have people build their characters. That way, at least, there’ll be no useless skills.