As I step into creating a die system, lets run through a few things I like and dislike, since these things will obviously impact the final product.
1. I want a reliable/predictable measure of the dice to be rolled. Anyone who has read this blog knows I love Cortex+, but my big complaint with it is that I need to pack _all_ my dice to play. Not because a given roll will use them all, but because I have no useful way to predict which dice I'm going to need in what numbers. This may be a small and petty thing, but there's a practical underpinning to it.
2. Next, I want to use d6s. There are some practical underpinning to this - they're ubiquitous and familiar - but it's also strongly aesthetic. I own lots of cool looking d6s that beg to be used.
3. I want the difference between skill levels (or whatever they end up being) to feel substantial, and I do not need more than 5 or 6 tiers of capability. This is a cinematic/fiction driven sensibility based on the fact that such broad distinctions make for solid character shorthands and are easily recognizable.
Up til #3, all of the options were on the table, but that last step there is going to make a count system problematic. Count systems may have very coarsely grained outcomes (based on number of successes) but the actual die pools tend to progress smoothly, with only moderate differences between pool sizes, especially at high levels. I could work around this limitation with something too-clever, but that seems like a peg-hole problem.
A flat system is still technically in contention, though it would probably require stepped bonuses. For example, my cold was game handles this by making skill bonuses (on a 3d6 roll) +2, +4 and +6. Those are a little close (they work better for Fudge) but the idea of stepped bonuses is not entirely off the table.
The tally system seems like the best contender, something in the Risus/Over The Edge/WEG space, with 5 levels ranging from 1d6 to 5d6 or something similar. Historically I might try starting from a baseline of 2d6 so there's a "step down" option and there's at least a little curve in the default roll, but I'm less attached to that idea than I have been in the past.
Now, there’s still nothing concrete to make a decision on, and this can be pretty paralyzing. Almost any choice can be made to work, so what do you do?
Simple: You do -something-. I’m going to go with a tally system because as cognizant as I am of it’s flaws, I’m even more aware of the dangers of sitting here waffling. So with that in mind, let’s see what we can do with a stack of d6s.
The first thing to do is to consider difficulties. I immediately rule out contested rolls because the last thing a tally system needs is more math, so that means fixed difficulties. Since I’m starting from 1d6 I think that means I’m going to pick the classic baseline of 4.
You see 4 show up in a lot of games. It’s a pretty convenient number for a bunch of reasons. On a straight d6 roll, 4+ means a 50% chance of success, and on a 2d6 scale it’s close enough to 75% to be reliable. On a range of die sizes it’s a number that can potentially be hit by a die of any size. All of which is to say that if you’re thinking going with 4 is a ripoff of anything, realize there’s a reason for its ubiquity.
Now, this raises an interesting question: if I’m allowing 5d6 to be rolled, is a base difficulty of 4 even faintly scalable? Certainly, the apex die pools should be reasonably rare, but that’s not any kind of excuse - a known, rare problem is still a problem. Thankfully, I have an instinct that makes this a little less problematic: I’m looking for success to be the expectation. Someone with 1d6 might have some trouble, and 2d6 still has some risk, but by the time you hit 3d6 it’s very nearly a sure thing.
That said, no reason to just leave it at that. Binary success is a little dull because it offers little differentiation between activities. Teaching high school physics ends up on par with crafting the theory of relativity. So that suggests to me that adding additional tiers of difficulty is the best solution.
The problem is that difficulty steps tend to be applied very arbitrary in play. Climbing this hill is this hard, but climbing that hill is that hard and so on. I want them to mean something a little more self-evident. And that, I think, is where I’ll pick it up tomorrow.