Sounds like a fascinating project—old tactics, modern math. If you can model the probability of each unit’s shift and feed that back into training, you’ll give a warrior a map of expected chaos. Just remember the battlefield isn’t a clean dataset; human unpredictability is the real variable. Keep iterating and you’ll turn discipline into almost a predictive science.