(Then she should go on playing for log-odds of success, right now, as she waits for that missing insight... this should not be said to Carissa at the moment.)
Has Carissa made any progress on finding new classes of highly stable and predictive magical dynamics? He has a computationally-complete set that he's trying to use for the current Magical Simulation of Magic; but he still has less total experience playing around with spells, and there's a proverb out of dath ilan: just because it's possible to compute anything in a computationally-complete-system doesn't make it easy.