Thanks, so if like DS 13-18 remain hotter than any other group for hundreds or even thousands of spins, that is enough bias to be profitable.

Its like the bias is hidden in the output numbers but if we reviewed the history of each number on a graph, we can see the prolonged dominance that is suggested by the output:

Here we can see only 3 hits over the closest rival in DS 16-21 and maybe only 5-10 hits over the rest of the board, but look at the expectation graph:

The red line represents expectation. So a clear dominance can be had with even small bias; this domination above expectation has persisted for about 150 spins (each red mark at the top is 5 spins and there about 30 of them).

I can consistently show this kind of dominance extending for many hundreds and even thousands of spins.

Sorry for the bug in the chart, this number hit so well it went out of range! (blush)

To further show the presence of hidden RNG Bias

^{TM}, here is the same session as it continued for the next 150 spins:

With only 10 hits above the median selection and only 3 hits above the nearest competitor, we clearly dominate the board and consistently provide hits above expectation.

The bias grows even greater over the next 200 spins:

Notice this is only 12 hits above the median value!

Again with the red line indicating expectation, this is a MASSIVE bias (106 LESS spins than expected for the coups taken) that has persisted for over 500 spins! How many spins do you play in a session?

Tracking the history is the key, statistical output will tell us what history to look at but

it doesn't tell us what happened spin by spin in the session.

Our mistake is glancing at a statistical output and going "oh, it all balanced out" --

IT DIDN'T!

Every hit above expectation represents 37 spins LESS for each of those hits!

So, this is what Dobble means when he says "a short run does not equal a long run" and "all of our sessions are short runs" and why Harry J says we "play the peaks and valleys":

......