Sorry for not getting back to this thread in a while; my response notification emails stopped coming, and I just happened to notice that I have replies.
My suggestion is don't do it via game design[...]
If the primary in game objective of adventures 1-4 is to kill orcs, you will likely wind up with a bunch of characters optimized to kill orcs, no matter how hard the game system discourages it.
If instead, adventure 1 is kill orcs, adventure 2 is steal something from the evil(or good, depending) wizard's council, adventure 3 is a diplomatic mission to a land where they abhor edged weapons and adventure 4 is to round up a large herd of horta, you will most likely have a rather diverse skill set in the characters when adventure 5 shows up.
While I certainly think that doing this would be helpful, I'm not quite sure how I could ensure this once I start selling my game on the market. Ultimately, the course an adventure takes is determined by the GM and players who are playing it; sure, I could design a few pre-made adventures to set an example, but that's about it.
What other game systems are you familiar with?
I'm mostly familiar with DnD 3.5 and 4e, with some experience in Pathfinder, Warhammer 40k roleplaying games, and World of Darkness. I also have many rulebooks for GURPS, though I only played one campaign of that, and it was a version which was both "lite" and outdated (even the books I have are newer than the ruleset of the campaign I played).
Though I loathe to play World of Darkness (I only played it because I had no other games going on at the time), I admit it's given me some ideas which I'm using, such as the mechanic by which your chance of success with any given action is determined by adding together two stats which do not directly influence each other.
My only suggestion is that you have the player flush out their characters some before they roll dice. That is to say, what is the characters drive, class, simple background. In the game I run when characters are being made the players have already decided what they are going to make for a character. And then they roll the dice. I must admit that the players in my group are all in the experience level of around 20 years of gaming different games and are more focused on making a good character than on the stats.
Yeah... some of my players haven't been
alive for 20 years, let alone having that much gaming experience. Also, since my game is new, and as such they don't have any prior experience with it, many of the characters they play are characters which I make for them, based on what they describe to me for what they want to play.
Also make them pay in game for those low stats. Whatever stat they have neglected make them use.
I'm actually starting to do this (remember the whole resolve thing I mentioned earlier), though I think it's sinking in for my players more because I'm giving them more info about what the stats do. Still, I'm having trouble with making a system with which to create a set of challenges (designing monsters, skill challenges, etc) which is both balanced against the PCs' ability and diverse in the type of challenges faced. Though I have tried many times to create such a system, most of the times I try fail to get beyond the "drawing board" or "brainstorming mechanics" stages, and so I end up making up challenges as the game goes along. I know, I shouldn't do that, but I can't think of anything better to do.