Interesting start, thanks. This definitely works as an initial concept. Limitation of current implementation being the reliance on a single geometry ID, but I bet I could work up some additional ways to work with the initial attractor concept.
The other nice to have would be the ability for the "objects input" in an effect to be a list of multiple instead of just one. Would be nice to have 3 different attractors for a single Forest object. Initially I duplicated the effect for a second one, which of course overrides the first. I suppose a different way would be to add the additional objects as more inputs in the single effect, and then figure out a different way to calculate the probability together. Hmmm - will think more on this, appreciate the first steps!