Things to Do

Dan and I have identified three main things that RPGs let you do: Tactical Combat, Problem Solving, and Story Telling. More on that in his post over at Geek Grab Bag (http://geekgrabbag.blogspot.com/2009/02/design-principles.html).

One issue with some games is that in each of the three situations, everyone must have something to do. This seems straightforward, but isn't really.

For example, in 1E AD&D, the first level magic-user does not participate in tactical combat (unless he knows Sleep, in which case, he participates for one round). Likewise the thief is usually at a significant disadvantage. Trading utility in one realm (combat) for effectiveness in another (out of combat problem solving, like our thief or wizard) is bad because it leads to players tuning out. Same thing with social skills in a skill system -- if you have 'em, then you participate a lot in role playing scenes. If you don't, then its safer not to open your mouth lest you botch that diplomacy or bluff check and ruin the whole party's plan.

4E did a good job with addressing this in regards to combat. For perhaps the first time, I feel like all characters have something to contribute in a battle. But there are still gaps in the other aspects of the game. For example, my rogue has little to contribute in the way of social skills and thus storytelling.

Since the OD&D Thief supplement in Greyhawk, there's been an assumption that trading raw combat power for utility skills is acceptable. I think its a false choice. I've been reading a lot about OD&D, pre-Thief class. Without a Thief class, ANYONE could participate in problem solving. Need to deal with a pit trap? The fighter can chop down a tree and put a bridge over it. Need to get past those poison needles? The cleric can ball up a rag and stuff it in the hole in the wall to plug it up.

Additionally, niche protection is all well and good. It encourages teamwork. However, sometimes its pointless. In OD&D, thieves, halflings, and elves were stealthy to some degree -- that was half the possible choices that could try to sneak around with some degree of success! Most parties would have at least two "sneaky" individuals. In AD&D and later editions, though, classes have become even more specialized. Thus, the hapless thief/rogue might be the only person that could sneak. How useful is it to sneak if you're the only one that can really try? That's just a bad idea.

However your system works, all characters have to be reasonably effective at Story Telling (engaging in the story, especially role play), Problem Solving, and Combat.

ADDENDUM:

In design, I think you need to carefully think about how the three-legged stool of Story Telling, Problem Solving, and Combat all interact. On one hand, you want shared mechanics as much as possible. You don't want to have to build three different characters with three different resolution systems for each aspect of the game, so some overlap is good.

On the other hand, the same mechanics that work at the tactical combat level may not work for problem solving.

I see at least two ways to think about it.

1) Use scale. Tactical issues are dealt with in terms of paces/seconds. Problem solving varies from paces/seconds (I need to leap across this gap right now to get away from this monster!), to furlongs/minutes (we need to find a way to bridge this chasm), or even up to larger scales (leagues/hours -- we need to navigate through the Enchanted Wood). Tactical Combat is traditionally used at the Pace/Seconds level, but one could also run it at the higher scales too, using larger numbers of monsters spread over larger distances with longer time spans.

You'd use the same mechanics at each scale level (so perhaps at the Pace/Second scale, a damage roll indicates HP done to one creature; at the furlong/minutes scale, a damage roll indicates # of creatures slain; at the leagues/hours level, its the number of enemy squads or platoons destroyed). The game zooms in or out as needed.

2) Use Aspect. So you have one consistent set of rules for dealing with Combat, and another for Problem Solving. This leads to overlap problems as we discussed above. You might need to use a Problem Solving mechanic in a tactical combat situation. 4E tries to integrate the two realms with Skill Challenges, but it doesn't really work. This approach requires you to fundamentally classify each encounter as a combat encounter, problem solving, or role playing opportunity so that you can let the primary sub-system dominate.