Thanks for the ideas! I've been feeling like the initial conditions concept could use some work, and this approach sounds promising -- seems like it could also

Thanks for the ideas! I've been feeling like the initial conditions concept could use some work, and this approach sounds promising -- seems like it could also facilitate playing around with different initial conditions in different plans (if desired) as well. I think warnings like you mention would be a great addition as well, though I have an intuition it may be difficult to pin down what outside of normal really is, since things can be different around the world. Still, I bet it's possible to call attention to any truly ridiculous-seeming values.