Hey.  I'll try to keep this relatively short.  I've noticed in RPGs there is this tendency for min-maxing, which is to say trying your best to be very, very strong in a few things, and running roughshod through the game, finding all the avenues that are in line with these strong attributes, and even finishing the game that way.

In Fallout 3, I remember Jeff saying he was good at-- conversation I think it was, and he got through the game that way, but ran into problems in the Alaska expansion which wasn't so conversation focused.  There is a real tendency, in RPGs where there's just one character you control, to allow you to maximize a certain stat, and then give you choices that seem optimized for your having really high percentages in that particular stat.  If you're good enough at something that you tend not to fail, it takes the game out of your hands.  There's little strategy in this.  If you keep picking head shots, and you're good at head shots, congratulations, you've secured yet another path to victory.  Maybe 2% of the time you miss, but that ain't often.

I got an uneasy feeling while watching the trailer for Obsidian's new spy RPG.  Will this be the same thing?  Get really good at pistols and kill everyone with pistols, even when it doesn't make much sense? 

RPGs, to me, tend to shine when you have a party, because a party allows these specializations to be spread out among several characters.  One character heals, one character sneaks around, one character blasts things, one character hacks things up into itty bitty little pieces.  This works because they watch each other's back, making sure any flaw in one character is balanced out by the whole. 

When in single-player games you should HAVE to change this strategy, otherwise the design is fundamentally flawed.  If you're a loner, you should need to diversify.  James Bond gets around because he has a series of convenient skills that get him from place to place (although why the guy isn't dead when he doesn't exactly hide is is beyond me).  If he was good at one thing, he might at best be a member of a group, or be inserted into specific situations where his specialty worked well. 

But unless there's a lot of richness in this specialty, in a game setting it makes for a very shallow experience.  Designers can't concentrate on one thing if there are several skills involved, otherwise these other skills are useless, so they try to make situations where you can use different skills to get by, but each of these are paths that lead to pretty much the same end.  A bunch of options, slightly different outcomes, but EACH of these outcomes in and of themselves are not terribly detailed.  Thus the only real fun is in choosing the path, and that pretty much begins early on, with the rest of the game charging forward to the conclusion (with hopefully some good, non-stat based puzzles along the way).  Unless the stats don't figure that heavily into it, it can't be like that if it's going to be an interesting game.

I'll probably add to this when my head's clearer, as I think RPGs have really gone off track lately and I want to make the point as concise as possible.