Unfortunately, the limitations on this discussion is neither the APs nor the casinos know the real numbers.
Sure, the casino has the money count at the beginning of the day and the end, but they still have to guess whether they’re only dealing with normal variance or some people playing advantage in the overall totals as to know just what the effect is on the bottom line.
Variance is huge for individual players but wouldn't getting a few thousand butts in seats bring it much closer to the norm.
Sure, the casino has the money count at the beginning of the day and the end, but they still have to guess
A few clever spreadsheet macros could tell them things with reasonable correlation, or at least suggest which small handful of players may merit further review.
There are likely several data patterns which are common to several types of advantage play.