Understanding variance

[Steve]

There's a ton of testing apparatus available, live wheels on sites, various forms of rng both on sites or imported spins into RX that can be live or rng, there's math....there's actual casino play results..

Yes plenty of choice. That's why you show parx & RS. And a few spins on RX. The point of MPR is its fair, realistic, and it is a good way to prove you arent just talk.

General pushed this as well because...well.... there's a reason.

Because MPR is a much better place to prove you're more than talk, instead of RS or parx. But you already lost on MPR, so you stay away from there now.

[falkor2k15]

Different variance test this time:

Each new repeat level shows how far out the variance happens to be from the LOTT expected (not the maths expected!).

This is different to measuring by spin #, as we can only tell variance change by the 2nd and 3rd place dozens (or if there's a new leader)

This doesn't include changes compared to the previous, as it's too complicated to code. Also, this shows "hit levels" specifically (as opposed to repeat levels) - and you need to subtract 1 from each result, as I couldn't work percentages using zero. Therefore, each cycle starts at 1 1 1, and then 1 2 1 = 1 unique; 1 3 1 = 1 repeat on dozen 2 (and 0 shows on dozens 1 and 3), etc.

It took quite a lot of data to find the expected LOTT values, and I wasn't able to comfortably go beyond 10 repeats. Now I need to see if there's actually any pattern here... 

[falkor2k15]

Just comparing dozens 2 (last place) and 3 (leader):

Repeat 1: Dozen 2 trailing behind LOTT expected by 50%
Repeat 2: Dozen 2 trailing behind LOTT expected by 67%
Repeat 3: Dozen 2 trailing behind LOTT expected by 75%
Repeat 4: Dozen 2 starts to close the gap (25%) - 1 extra hit over dozen 3
Repeat 5: Dozen 2 goes back out again (40%), but not as much as before (75%) - dozen 3 had extra hit
Repeat 6: Dozen 2 stays out at 40% - dozen 3 still got the extra hit
Repeat 7: Dozen 2 goes further out at 43% - dozen 3 and dozen 2 both 1 hit
Repeat 8: Dozen 2 stays out at 43% - dozen 3 had extra hit
Repeat 9: Dozen 2 starts to come in (20%) and suddenly has 3 extra hits over dozen 3!

Dozen 1 looked quite random - in and out constantly! This variance is hard to tame....

But maybe there are extremes:

Dozen 1 suddenly had 6 hits right at the end!