I've always appreciated your posts. I find them well thought out, curious, educated, and rational. OOTP has had significant input from ML employees I believe. One of the MANY reasons I wanted to hire an employee within the analytics department was his involvement in OOTP from both the technical side and his involvement as a player. I thought there were some things we might be able to explore, verify, and implement. In addition, I agree with you on analytics being more than formulas. I would say on a foundational level it's helping you sort into "red" vs "green" as you mentioned. Just starting here is a necessity. Analytics should also help you ask better/deeper questions and lessen risk.
OOTP or any other well-constructed simulation exercise, should give you insight into risk, and also help maximize your roster construction. These simulations should aid in maximizing platoon situations and positional flexibility along with bullpen chaining.
As you mentioned, the acquisitions of those 5 pitchers together is confusing. All have injury and/or performance questions attached to them that don't play well together. I agree with you, building a team is like building an asset portfolio. You're trying to balance asset allocation, sectors, cost, risk, etc....Perhaps it wasn't the strategy, but the execution or difficulty in executing the strategy. I'm not sure.
In the end, I do agree with you and appreciate your insight and ability to mix in "real world" experience into the game of baseball.