I wonder how far you can get by studying high speed video of pitchers, their arm angles, arm speed, release point, grip, finger measurements etc.? There are so many factors to consider but I wonder if some thorough study would be able to get to a point where you have a well-defined "expected spin" for a pitcher's repertoire, and any level of spin above a certain point is flagged as suspicious.
Maybe impossible in practice but it's interesting to think about.