Thanks Cyangmou, unfortunately I came to the same conclusion that my reference pick was a poor one, but a little too late. The second one I posted was me trying to deviate from it a little.
I agree that this would be theoretically easier to do at a higher resolution, but for me (for now) I prefer working smaller because it feels to me as though rather than having full control I can hide some of my incompetence in the sprite size.
The best I can compare it to is guessing on a multiple choice question, if there's only 4 choices you have a 25% change of putting the pixel in the right spot, but with 8 choices I have more freedom than I know how to work with.
I personally prefer the second skull, though that's coming from less experience than you. One of the biggest changes I made were the teeth, as I'm mostly doing this to practice that "cluster" concept and wanted to find a way to forego the isolated pixels.