It's rare though for some huge breakthrough in efficiency to happen.
That's often the case. I'm working on a project right now to try and improve certain outcomes (I'm being purposely vague here) that, after a couple of more months of research and model building, will probably be scrapped so the firm can stay with their current model (coin flip).
It's entirely possible the M&P that's currently doing it also do other "value-added" things for MLB as well, but that's impossible to know without more information.