I'm returning to my Game Chef 2011 entry, to rework and polish it and, you know, turn it into a complete game. One of the things I want to do is play with the dice mechanics, see if I can come up with something better--more flavorful, self-balancing, and/or less mathy.

To review, the current mechanics are d8 dicepool, total up the dice, and compare to 5x Titania & Oberon's stats (separately). The upside of adding the dice, rather than counting individual successes, is it makes it possible to succeed at a task when the sovereign's stat exceeds your score. The downside is math, even if it is just adding up a handful of single-digit numbers. Also, knowing how bell curves operate, the more dice, the less the odds of exceeding your average are, even if the nominal range widens. So I might just be fooling myself--and, more importantly, players--into thinking they have a chance when, realistically, they may as well not.

The other problem, which isn't specific to an additive dicepool, is that the game is too loose: it's way too easy to always use your best score. Now, some of that can be addressed narratively, but I don't want to make it a game of outsmarting between GM and players. It's meant to be more of a storytelling game, with the GM just framing the story and providing an opportunity for player cleverness (or at least entertaining one another). But I'd love to tweak the mechanics such that there are mechanically-inherent reasons to not always use your highest score. If I can build that into the raw dice mechanic, great! I suspect some sort of resource management layer may be necessary.

Now, one possibility is a d8 dice pool, with the difficulty equalling the skill. So, if you're rolling 1 die, every roll is a success. If you're rolling 5 dice, rolls of 5 or better on each die are successes. Now, my gut is that the odds of getting at least 1 success will stay roughly the same--close to 100%. But it has to go down at least a bit--it's possible to fail if the difficulty is 2 or higher. Plus, it becomes possible to get more and more successes, even if the odds aren't awesome.
Conveniently, the game already caps scores at 8, so you always have at least some chance to succeed.

However, I'd like a bit more detail on the dice behavior. Anybody able to help? Or point me at a dice odds calculating tool that handles individual-dice successes like this?

Ideally, the odds of failure would climb steadily with increasing skill, but not get too high, meaning that it's a gamble to use your higher scores: more likely to fail, but a chance of getting lots of successes. I want there to be a mechanical reason to not always use your best score, because it's a very freeform game where you can probably narratively contrive an excuse to use your best score every time.

-Nat

Pardon me a bit for rambling on sight unseen, but it does sound like you're trying to punish the players for wanting a skilled character.  If your design goal is to get them to use a lesser value, perhaps you can bribe them to do it themselves.  How about having the players burn dice from their pool in exchange for some narrative control? 

Sure, anyone can kill the guard with 5d8, but can you get him to spill the beans about the castle defenses (-1d8) and the location of Baron Von Badass (-1d8) in the middle of a sword fight (now 3d8)?
--
TAZ 

Not so much punish for being skilled as reward for being diverse.

But even that's not quite on point. It's intended to be--and currently is--a pretty freeform game, where the faeries' powers are very broadly defined, and the immediate goals are very much player driven. As such, if there's no incentive to use something other than your best power, a player could probably always contrive to use their best power. Which is boring. Since the rules and situations are unlikely to ever force you into using a weaker power, I was thinking to provide a mechanical incentive to do so--some reason why, in some circumstances, it would be preferable to use your weaker powers. So I thought that having less chance of total failure might provide that incentive, provided that there are varying difficulties so sometimes you would need to take the risk of a higher chance of failure, in return for having a chance at achieving that higher success.  But if you either know or suspect that the difficulty is low, using your weaker power would be the best way to go.

Or, at least, that's my goal. I'm not sure if what I suggested does that.

-Nat