See? This is why it's good (but painful) to sometimes expose one's ideas to the 'Net.
All the posters who commented on my last column made very good points. As mentioned in a response comment, the original idea for the mechanic came from Silhouette's basic skill test mechanic, which involved rolling a number of skill dice, adding an attribute-based mod to the highest die, and then comparing the two highest dice and using the remainder to determine the effects. I liked the idea, but started tinkering with it and looking for ways to make it my own. What ultimately came about was something that, apparently, is a non-starter.
Which leads me to a point I want to emphasize, both to myself and to my readers who are also considering work on their own systems or rules - don't be afraid to kick the baby.
Somewhere floating out there I once came across a list of 10 (?) tips for indie RPG designers (here's a list of good ones, although this isn't the list I'm talking about). Maybe it was written by Ron Edwards, maybe Robin Laws - I can't remember. But one of those points was that you cannot cling to one concept so tightly that it blinds you to the fact that it was never worth holding on to in the first place. For a lot of people (myself included) RPGs are a very personal, creative thing, and a lot of us take great pains to shield our creations from the harsh glare of objective criticism. In the end, as with real children, if you shelter it too much, your creation will suffer.
[Edit: My good buddy Masakari was able to find the post I was thinking of. I was thinking of points 2, 3, and especially #10.]
So, I'm tossing aside my High Roll mechanic. I actually think I'm going to use a more Traveller-like idea; 2d6 + Skill Rating, aiming to beat a base difficulty of 7 (so 8+ is a success). It's quick, it's (I think) pretty straightforward, and for a lot of people, it will be at least passingly familiar.
What do people think - an improvement? Comments and questions are, as always, welcome.