After one too many games of Guess The Key at night, I finally decided it was too much to have both an ignition key and top box key on my ring: I want to be one of the cool kids and have a single key for everything too!
I started by following the instructions at [How To: Re-Key Your Top Case], but at about step 5 I realized that those springs and wafers are like really really small and I don't have the patience for bettering myself on this particular topic right now at this table. But what to do?
Since I had bought a new lock to work with just in case I managed to massively screw things up with all the teeny bits, I decided there was no loss in exploring the destructive route. Since all wafers must be in for the plug to rotate, I simply inserted the ignition key and found that only three wafers remained up enough to prevent rotation. Target acquired.
Talking to the nice guy at the hardware store about what tools I had and what I wanted to do, we decided that I should try a chainsaw sharpening stone for my dremel and a vice grip to hold it in place. Trying to grind down the plates was time-consuming, but possible. During the process, I twice tried to keep the wafers from jiggling by jamming toothpicks into the cracks between them and the plug, but to little success...until those two wafers just popped out during one of the turn tests. The first one had already been ground down sufficiently with the grinding, so off to the races.
If I had to do this again, I would just try and figure out how to eject any blocking wafers upfront and not even think about the grinding (even though it was eventually effective).
Reassembly was easy (I took pictures as I went and reversed through them) and I was able to soon confirm that I had "re-keyed" the top box. Success! But at what cost?
I'd be interested knowing what the real numbers are, as I'm a little hazy on the internals, but I believe that the way the top box lock and key works, there are only two possible states per wafer for the original keys. (I'd note here that the ignition key seemed a little more subtle than the top box one; maybe tri-state? It doesn't matter in the end on the top box side.)
Given that there are two states per wafer and nine wafers, there are 512 possible states (expt 2 9); a random key in the same profile group would have a 1 in 512 chance of unlocking.
So how much destruction? We are left with 6 wafers, giving (expt 2 6) = 64 states, which is (expt 2 3) = 8 times less "secure" than the original. For my particular use case, almost any disjointedness with the ignition key would have given fine results, as I just want it closed and locked to a tug and the eye--I'm not worried about a lockpick going to town. If there had been no disjointedness, I would have been the lucky winner of the ignition key and top case key being naturally aligned; if there had been full disjointedness (1 in 512 chance?), all wafers would have needed to be destroyed, rendering the lock pretty much just a visual deterrent.
I think that proves that almost anybody either already has a matching set or can make a set with minimal effort that's at least "minimally secure" (i.e. 50/50 chance with same key profile).