TurboGenius Coldest Number Theory

Simplified down lately (for me) just using the even money and dozen/columns to "predict" a number.
Same results though, many wins within just a few spins at a rate that doesn't happen with just random
numbers being picked.
I'm sure the math people say otherwise.
I also want to start as early as possible - tracking 10 spins and then making the prediction
based on the even money and 2:1 bet locations seems to work just fine - but not all the time
of course.

 

Nice, that takes care of using pretty much all the time to complement it with so-called hot.

Still, I only need 3 numbers per each 18-area (2x 3SU=ST), so I would need to sieve'em out.
For the times it works, & for those it doesn't work (I presume you'd progress in risk);
the credit points applied then are the 'difference-1' correct?


Ongoing 3-number determination, when necessary only;
I would take the active 3x-peaters, perhaps even 4x-
from current or previous 37-spin cycle,
then complement'em with the hot counterparts ..

.. but on a 10-spin rolling or similar it would take a while till ≈spin100
weeding out the unhit remaining, meaning it would leave out a lot of unhits
to opt from relevant to only 1-2, occasionally 3 Sú numbers required.



Alternatively, I could simply play a ST; so require to have a ST ready for each side.
Any other advice?



I've sent you an email on your gmail account for context ..
links to two RS games - one 3SU option, the other ST option.
Have a quick glance.

 

Interesting, that's what so great about doing the manual work on your base premise, because enough spins are evaluated it will be easy to analyze the results in multiple ways to find the an optimal risk/reward/time spent approach.

Things like:

Is the hit rate/prediction accuracy higher if tracking is done on 10, 20, or 30+ spins?

Is the tracking length more, or less, important than the rules of the scoring system being looser (more numbers) or tighter (less numbers)?

How many betting attempts are optimal once the predicted numbers have been determined?

Those blanks can be filled in (within reason) if the sample size tested is large enough, along with validating or invalidating the core concept along the way.

I like it already tbh, it's a very cool twist on the usual binary and limited cold/hot selection debates.