Game Mechanic Concept...

[Clint]

+ Subtract CS from Luck Roll
+ Compare CS - Luck Roll

First off, the flip-flop in terminology needs to be fixed.  One step says subtract the CS from the Luck roll, and then the next says subtract the Luck roll from CS.  That's two different things.

Second, what's the range for modifiers due to attributes and skills?  At this point, it must at least run from some theoretical negative value to 28.  Is that it?

If so, then with a negative CS, it would end up being added to the Luck roll.  Note that this is based on the statement that a roll equal to or less than Fate is a success.  The fail then on a Very Easy task would be to add a number to the result of 4d6, which has a maximum value of 24.  Thus a negative CS possibility is the only way to fail a Very Easy task.

This then gets into the very confusing game mechanic of a "double-negative" or subtracting a negative value.  There are other ways to achieve the same mathematical effect without going that route.

Ultimately though, I have to say that the subtraction mechanic and negative to positive scale is
confusing and complicated, and using it solely for the sake of being "different" and "unique" is pointless because it is neither.  It is simply another mathematical formula for the same basic mechanic.  If you are going to use that mechanic, then the easiest version of it to understand is obviously the best one to use (die roll plus modifiers equals or exceeds difficulty level).

The real key to being different is to change the mechanic not the math.  Otherwise, just use what works; it's not like d20 invented the concept.

[M. J. Young]

A lot of how this plays depends on the values you're allowing for Ability and Skill, but the curve is already rough.

4d6 provides 6x6x6x6=1296 permutations, of which exactly 1 will be all ones and the same all sixes. So your extremes appear a bit less than one time in a thousand. The next step, 5 or 23, each appear 4 times in 1296, or once in 324 (I'm doing these numbers in my head, but if they're wrong, they're close), which added to the extreme means 5 chances to roll not greater than 5 (which of course is the more desireable roll), 1/259, roughly. Permutations get very complicated after that, but generally the more dice you have in play the more tightly your results will tend toward the center point--in this case, rolls of 14 will be extremely common.

The question then is whether you want a randomizer that is so focused on the center.

If you're looking for ways to generate curves with 24 at the high end, 3d8 is considerably gentler (512 permutations), and 2d12 is pretty much straight up and straight back down again, starting with 1/144 at the extremes (2 and 24) and climbing to a peak of 13/144 at the centerpoint (13).

I know there are people here with software to calculate the permutations quickly; I don't have that right now, but I really do have trouble with a 4d6 curve if you're not trying really hard to minimize the impact of the randomizer.

--M. J. Young