Hey again everybody!  I had an idea for the core mechanic of a secondary game design that I'm working on, and I thought I'd post it here and see if it stands up to scrutiny.

To sum the game up quickly; it's an anime & video game inspired swashbuckling / martial arts fantasy game with steampunk elements (kind of Final Fantasy meets Soul Calibur meets Studio Ghibli).  I haven't settled on a premise as such, but I'm aiming for a mix of sim / narrativist play using fortune in the middle, and motivational / emotional factors.

I'm looking at a centralized resolution mechanic using polyhedral dice.  Originally I was planning to rate skills as die types / sizes (D4 to D12 or D20), and rate difficulties the same way, so that a standard roll would involve two opposing dice, creating a triangular distribution skewed toward the higher die type.  Degree of success would depend on the numerical difference between the two rolls.  The problem is that I had trouble accounting for bonuses & penalties, since there were only 5 or at most 6 steps using standard polyhedral dice.

So after a lot of banging my head against a table, I came up with the following last night:

Abilities and difficulty thresholds are rated according to die type / size; and a second factor for each roll (either Effort, for the player; or Resistance, for the GM's opposing roll) indicates the number of dice rolled.  So, Ability D10 and Effort 2 means roll two ten-siders.  Whoever has the highest single die showing wins the contest.

Effort dice are purchased using an action point mechanic as currency... so you can spend a 1-4 action points and roll 1-4 dice, depending on how committed you are to the action.  Your action points are determined using narrative character qualities such as motivation, moral attributes, gimmicks, etc.

Resistance indicates the complicating factors over and above the base difficulty threshold for the task...  so, shooting a stationary target at medium range might have a threshold of D8 and resistance 1 (i.e. 1D8).  If the target is moving, the resistance increases to 2 (so roll 2D8).  If it's also dark outside, resistance increases to 3, and so on.

Degree of success is measured by the number of dice that beat the highest opposing die.  If 1 die beats the opponent's highest die, you have a partial success.  2 dice = good success, 3 dice = critical success, etc.  Anything else is either a failure, or a tie (if your highest equals the opponent's highest), and is narrated accordingly.

So, the question is: is this way too complicated?  The original mechanic was intended to be fast and dirty, a quick roll and you're off to narration.  This new method seems interesting, but seems like it takes a lot of time to figure out all the factors involved.

I have some alternative methods in mind, but they seem to involve using pools of all the same die type.  I'd prefer to use a variety of polyhedrals for aesthetic reasons, but I'm not certain if that's feasible.

Pick your die size. Pick your effort. Roll dice (just a few). Compare highest value with opponent's. Tie? Done. Opponent wins? Done. You win? Check how many dice beat opponent. Done.

Not complicated at all.

Does this mechanic give you the statistics you want?

It seems to give me good stats.  A range of 5 or 6 die types for ability scores; and three increasingly unlikely degrees of success (as well as failures and ties). Modifiers can be handled by affecting the action point pool from which the Effort dice are drawn, while situational complications change the resistance of an uncontested roll. 

I suppose there's a bit of a problem if there are situational complications to a directly contested roll...  for example, Alfred is swinging a sword at Betty, who is trying to parry.  Let's say there's some sort of an earthquake happening, and Alfred is off-balance; does the penalty affect his action points, or does it add to the effort of Betty's roll?  I guess in a situation like that, they're probably both affected, so the penalty cancels out, but it's still a bit of a question.  I think it's a bigger deal for ranged attacks, where lighting, movement, and cover can affect the shooter's roll but not the target's...  but then ranged attacks don't necessarily have to be contested rolls--dodging may just add to the resistance.

Still a bit to think about...