Well Stepy, it is "showing form", isn't it? Kudos and thanks for sharing your tests with us.

Remember they do not have to be exactly like the dozens/columns! These cadences may have their own patterns which is okay, what you are looking for is consistency, events you can spot and which are somehow clumping and frequent enough to lay your chips on them.

Yesterday I comented with dealer friend, Lanky's wheel is located in Australia and he plays here at Venezuela, yet the same Lw patterns work as they are using roulette's true nature and therefore they are universal. Now that's consistency! 

Regards,
Victor

Hello all

For those who wonder what happened next, here is my conclusion : don't try this.  It is not efficient, and its basis is wrong.  Here are the two reasons why :

• First there is a conceptual mistake : my starting point was to put a three part event into 2 possibilities, like the original "Lw" on dozens-columns.  I turned this into "did the ball land 2 times in a row in the same cadence or not" ? This is wrong thinking.  The original idea behind "Lw" is "did the ball hit either one of the last two dozens".  So to incorporate the "48" cadence into the remaining two was but the continuation of this misconception.
  I would had better said : "ok, consider the 3 cadences as 3 groups, then ask : did the ball hit any of the last 2 groups", so no need to get rid of the "48".
• Second there is a beginner mistake (the beginner being myself) : insufficient testing.   At first sight, it seemed I had found a good representation of how the game was behaving, because my "Lw" tracking of the 2 - 3 first series were encouraging.   But further tests proved reality did not follow my system, (much) more often than it did.  In addition, the "Lw" original rules could not apply, there were no consistency.  I. e you have three "Lw"s, you are happy, you bet next "L" and boom here comes chaos : LLLLL, or anything different than the expected "Lw".

In statistics, there is a theorem that states you can approximate a binomial distribution by a normal one when "n" goes past 50.  I did not even test my ideas on 10 permanences before posting ! So forget about all laws, and of course accuracy of analysis !

There is good in all this, though.  As for me, it helped a lot thinking about roulette.  Indeed, putting my ideas down into words got them out of my head, forced me to formalize them, and to see clearly where it was wrong afterwards.  No writing, no seeing (or at least, nearly not as clearly).
So I pushed the analysing process further, but this would be the subject for another thread.

Admin and Mods, if you want to pull this thread down in order to save space or time for future readers, I certainly would not mind.

Cheers, 

Stepydan.

P. S : many thanks to all contributors for their time and efforts.