the unknowns that remains are around how much all the physics metrics get to the last, hardest part of hitting as a player levels up to the top tiers, and that is the tougher decision making. i.e., Your ability to barrel up more balls better than the next guy against AA pitching isn't necessarily 100% indicative of your ability to see and react to spin at MLB command and velo levels. Recognizing better disguised pitches and decreased processing time are parts that remain as hurdles that are harder to project from lower level data. That - and of course injuries, are residual uncertainties even with the tech.
To me the irony is that the better metrics do tell you almost everything about a pitcher today - those more subtle vision/processing issues don't apply to their art, but the injury risk with pitchers is so high today that the increased certainty about their stuff still leaves huge overall uncertainty there as well. 🤷♂️