Ok, so we now have a basic die mechanic (a tally system for dice pools ranging from 1d6 to maybe 5d6) and a basic method for establishing difficulties (numeric steps 4/7/13/19 based on expertise required to perform a task) and a methodology for handling success. That's a good start, but for all the thinking that went into it, there's not much there. In the absence of further details, this is just one more d6 system looking for an answer to the question of "Why not just use RISUS?"
First off, let me state that RISUS is a good answer to many questions, and if the sole purpose of this were to create a not-RISUS, that would be a little lame. There is a lot of desire in RPGS to differentiate purely for the sake of differentiation (or, more cynically, to make something different enough that you can put your name on it) so beware that particular lure.
Thankfully, I have enough things I want the system to accomplish that we still have some very organic work to do. Let's start looking at the bells and whistles.
When looking to add fiddly bits to a system, its worth thinking about when in the process you want things to happen. Are you going to introduce choices before the dice roll or afterwards? Or both? This gets further complicated when you have more than one roll (such as D&D's attack & damage) but let's not borrow trouble. We'll assume a single roll for the time being.
When making this decision it's worth realizing that there's more than just timing at work. Choices made before the roll have a very different texture than those made afterwards, and the difference revolves around the certainty of the outcome. Any choice made before the roll is, practically, a gamble. It's a choice being made on the hope that the subsequent roll will succeed. This has the benefit of being more engaging and the drawback of risking paralysis - it's easy to get hung up on what would be the "right" choice.
In contrast, choices made after the roll are very certain, and tend to be all about working with the outcome. In addition to being simpler, these choices usually are much easier to ingrate into the fiction of play. That is, characters don't use their big attack to miss - they hit, then decide it's going to be a big hit. For some people that logic is jarring, especially if the game has enough missing that it's expected, but other players find the other disconnect jarring.
So, let's say we want a before-the-roll mechanic. We probably want to embrace the gambling element of it, allowing the player to take on some risk before the roll in return for a potential reward. The easiest way to do this with the system so far is to play around with the trickle-down nature of success. Under normal circumstances, even if you don't hit the target you're going for, you still succeed to a lesser extent. You could offer a bonus on the outcome provided that failure will be absolute.
That feels a little awkward, and there are simpler alternatives. Bonus Dice, for example, are a common trick for rounding out dice pool systems. That is to say, you may get to roll extra dice, but only count the best ones. These might be true bonuses (Like in PDQ# or Over the Edge), fixed result pools (like Cortex+ always keeping the best two dice) or roll & keep systems (like L5R).
This is a pretty robust system, and has the advantage of keeping the results within the bounds of the original die pool. It's worth keeping in mind, but it's also well worn territory. I'll keep it in mind, but I'm not particularly inspired. With that in mind, I'm switching the focus to an idea for an after-the-roll bonus.
A while back, a very clever game called Secret of Zir'an failed as a result of a terrible printing failure. One of the more fun ideas it included was a very robust system for using the Margin of Success on a roll for mechanical purposes. That is, if you needed a 20 and rolled a 25, you might spend 2 points to do extra damage and 3 points to knock your opponent over. Neat little mechanic, and one I'd love to capture it.
The problem is that margin of success math is a pain in the ass. Adding up a bunch of dice is enough of a barrier - adding a second equation to the mix is just a bad idea, so the trick is to find a way to make that simple.
With that in mind, we'll try a trick that steals a page from the idea of bonus dice, but keeps the focus on the margin of success. The player may "keep" however many dice he needs to make the roll - if there are dice left over, then those can be spent as currency. It doesn't matter what the values on the extra dice are - the dice themselves are currency.
For example - let's say you're rolling 4d6 trying to hit a 7. You roll 2, 2, 3, 4. You can build the 7 out of the 3 and 4, with 2 dice leftover. Those two extra dice can be spent for extra effects (akin to Dragon Age's stunt points). I don't yet know what the points will be spent _on_, but I think that gives enough of a starting point to start hanging some bits off of.
Now, it's important to remember that these bonuses need to be tangential, not additive. That is, no number of bonuses should be able to turn a 4 result into a 7, which means the bonuses don't make the result _better_, they add additional elements to it. In the most broadly narrative sense, each bonus might be used to declare a fact of some sort. This is a fuzzy distinction, but it gets concrete when you start building specific rules - it's easy to look at any specific bonus and compare it to what a higher roll would have done, and see if there's any overlap.
Curiously, this little test also sheds a lot of light on what needs to be thought about in a conflict system, so we'll probably head that way next.