Reyth, lm confused beyond my capabilities to understand.

There are levels of abstraction that are conventional: average , std, chi- square, level of confidence ( chance of randomness)...

There are not " conventional " ones like coefficient of stability, skips/ repetitive hits likehood...ets.

We need these to determine how likely phenomenon that we observe is to repeat itself in play. It all boils down to the correlation and stability of correlation, if you wish " index of correlation coefficient "... all of these are valid math tools that show how likely and stable is phenomenon that we observe.

In the nutshell, we have to have a " trigger" that says " bet there when..." and it should be statistically significant. If not, what we observe are just flick of fluctuations that we cannot predict, be it rng or real wheel.

Naturally we are interested only on phenomenon that is consistent and predictable. .... if idea is to make money of cause....