Before i start referring to playing models from the book i want to mention this quote.

"IS EVERY ROULETTE SPIN NEW?"
Marigny de Grilleau
translated from "The gain of one unit on the even money chances at Roulette and Trente et Quarante"


One can hear that question in every casino everyday.
The word "new" means according to the definition "which one yet did not see".
In this sense each day is a new day.
It is quite obvious that people asking this question do not realy mean "new" to express this natural truth.
Their questions is badly formulated and surely they mean "new" in the sense of independent.
Thus they wanted to ask whether each spin is independent of the others, the previous or following spins.
The above question should be asked as follows: " Are all appearances and are all spins independent?" In this formulation no wordplay and no wrong interpretations are possible.

Grilleau does not hesitate with a clear answer: "No, neither the appearances nor the spins can be independent, because everyone of them is a part of the whole. This whole is arranged and limited in all its movements and is subject to precise laws."

Each spin, while the ball turns in the wheel, carries in itself a certain quantity of independence and a certain quantity of dependence.

The independence results from the following:
every time the dealer rolls the ball, it is faced with 18 red and 18 black, 18 even and 18 odd as well as 18 high and 18 low pockets. Therefore the ball has the same chance to fall in one of the 36 pockets (we do not consider zero or doublezero this time) since each pocket indicates Red or Black, Even or Odd, High or Low at the same time.

The dependence results from
1. the Law of Deviation (Ecart),
2. the Law of Balance (Equilibrium)and
3. the law of the distribution of appearances into different accumulations or clusters and isolated units

Thus the mathematical truth of the independence of the spins is constantly in conflict with the statistic truth of the dependence of the spins.

If between two equivalent appearances none, or only a very small deviation exists, the independence of the two appearances remains retained in their fight against each other.
But if the statistic deviation reaches a certain size, the size of this deviation more or less limits the independece of these appearances and spins.
In this instant the dependence of the appearances on the laws of nature demands again its right, by limiting its freedom for deviation within the statistic average values, of which these never can free itself.

In our opinion neither a single spin nor an appearance can be independ in a roulette permanence of a certain length, for example within 1024 spins.
The dependence of the spins which are affected by chance due to exactly defined laws, is a fact, which the usual gambler does not understand without difficulty. And because of this difficulty the gamblers and also the mathematicians believe in the independence of roulette spins.
In reality each spin and each appearance has its necessary and mandatory function in the whole of a roulette permanence.
Chance does not exist there, because all effects have their visible or hidden causes.


The dependence of the spins on the laws of nature becomes obvious, if we analyze a roulette permanence and classify the developed appearances.
However we do not succeed in each case in determining this dependence, which must be present for all spins, if only small deviations occur, which do not exceed the average statistical Ecart of 1.
We only succeed then, if we determine the partial return to equilibrium after very strong deviations greater than a statistical Ecart of 3.
The roulette ball cannot extract itself from the laws of nature.
These laws force it into the pocket, into which it must fall, so that it can perform the necessary function, which it has to, in the statistic harmony of the whole permanence - like a note in a score.
Chance can let many obvious, strange features develop before our eyes. But nevertheless, statistically seen, chance can not repeat these individual strange things too frequently, like for example a series of 25, which needs approximately 34 million spins to develop once.

Its based upon my understanding about the work in general.
I start from this position to solve the core about the subject.

First we do is using some kind of playing model.
With this example i will referring to singles contra larger series.

We look for imbalance or bias within a window of events / trails.
This means we are measuring overrepresented events and underrepresented events.

When we reach a benchmark around 3.0 STD (could be any) then we have a pretty strong imbalance / bias.
Then it can continue to grow stronger as the nothing is due or that the bell curve has no limit.
This does not effect us.

This window of bias or overrepresented events following strict rules and values based upon math and probability to be valid.
With the example singles contra larger series so do we have the following values or playing model.

Singles has the value of 1
Series of two has the value of 0
Series of three has the value of 1
Series of four has the value of 2
Series of five has the value of 3
Series of six has the value of 4
Series of seven has the value of 5
And so it continues ...

Look at this values again, here you have singles contra series of three and both with the same value.
This is because when you measuring the random flow or distribution of random events you want to make sure you have overrepresented events and underrepresented events or a true bias into one direction.

This way with this example larger series are underrepresented.
Now you can pick how many windows you want with 14/2 and you will for the next 50 100 150 200 250 300 trails get larger series.
You know they will show, but you don't know when, this is why cut point methodology only attack when change is present.
More about that later.

Example 14 singles and 2 series of three is 3.0 STD.
it's a window of 16 events with a strong bias or a strong window with overrepresented events.

This is the starting point or core about this subject.
To measuring the random flow and find overrepresented events or strong bias (around 3.0 std)
Different methods or measuring other events has other values based upon math and probability.

With this example we have find and observe a window with 14 singles and 2 series of three.
From this moment and further into the future so do we have some degree of expectation.

I am not saying we have a crystal ball.
But we have what we can call vacuum pressure or a hole or a window with strong imbalance of overrepresented events.
So our expectation is that sooner or later the underrepresented events will start to show.

I know that the bias can grow and get stronger and hit 4 5 6 STD but it does not effect us.
In this case we would only observe.

Cut point methodology is about to capture the drop point after a strong bias.
The drop point or the underrepresented events can come as tiny, medium or large drop points.

The main idea is to capture them when they already show or are present.
That way you have a indication that the bias or imbalance has stop growing stronger.

This is what i call tendency play.
We can also see it as regression towards the mean or correction.

It is the best existing trending method for even money bets.
Based upon pure math and probability.

The key to solve the final solution is how you define a march to capture does events with money management.
Measuring and find bias is the easy part.

This is how it looks like with the example i mention above.
Singles contra larger series.

2 x
1 x
2 x
1
1
1
1
2 x
1 x
2 x
1 x
2 x
1 x
2 x
1
1
2 x
1 x Statistical Ecart around 2.5
2
2
2 Stop growing stronger and getting weaker.
1
2
2
2
2
2 Larger series
1
2
1
2
2
2
2
2
2 Larger series
1
2
2
2
2
2
2 Larger series
1
2

The key to the final solution - Marigny de Grilleau
Is a book from a Swedish author (Sten Nordland) and the main titel of the book is International Roulette.
Its two books.
Part two is exclusive about basics and advance new development about Marigny de Grilleau's work.

I will write about this to become a reference area for cut point methodology or also known as regression towards the mean.
This will based upon my understanding about the subject.

Quote from: Bally6354 on November 25, 2013, 07:51:57 PM
This is where it unravels.


[attachimg=1]


-19 in these 8 games.

A few double losses and wins on the third attempt clustering together will put you in a hole using a progression. It's bad enough flat betting.

So I am done with this one.

A nice try..... no 7% edge flat betting as claimed IMO.

Bally's rating......4/10. (There are a few good anecdotes in the book)

I think you give up to easy.
-19 in 8 games is your last post, but before that you made 12 session +25 units.
That add up as with any good method there will be bad days and fluctuation.

One free site where you can test your system for free www.psroulette.com

Cheers

Thanks for the link and i know random org, i use it all the time :-)

I forgot where i can download actuals from real casino ....
If you have a link ,,, many thanks ...

Cheers

Quote from: Mike on November 21, 2013, 05:33:11 PM

They should do, but I'll get back to you on that...

Sound great.
Mr Ops site is gone now, but he did a test using the past 6 and got some surprising good results.
This was because a user tell him about it and that it was the only method he had some success with.
Then Mr Ops got curios and run some simulations.

Lets say i am not want to play this method to win money, i just want to break even and go home with my collected data.
So the main idea is to use a method where you can walk in and walk out with 300 trails collected data about the wheel.

Or a method where you have a entering point and one exit point.
The method using skips.

+ + - + + - + + - - + + + + - + - + + + - + - + + - - + - - - +

Lets assume i would play once after a loss up to two attempts and we can name it march 1.
Then the result would look like this W W LW W W W W LW LL

Lets assume i would play once after a win up to two attempts and we can name it march 2.
Then the result would look like this W W W W LW LW LW LL

Looks pretty much as even money for me or with this examples you would probably end up with profit.

Nice Turner ... can you give some example ...
If i understood it correct you see 12 numbers with no repeat and then bet the following 12 ,,, is that correct ?

I wounder as this topic is about the "periodicity window" where repeats appears at most or most common.

I am not sure, but with simple observation you notice clustering periodicity.
You get repeats with in a window frame.

Can be 6/6 or 5/5 ...

I run 300 trails with random org and got the following results.
+ sign means you got a repeat for the next window of 5 attempts.
- sign means you did not get a repeat for the next window of 5 attempts.

+ + - + + - + + - - + + + + - + - + + + - + - + + - - + - - - +

19 windows where one repeat appear
13 windows where one repeat don't appear

@ Mike are you saying the windows even out in the long run.
I was thinking the state would be more common then the other ( periodicity ).

Assume you track a wheel collecting data, then this look like you would not lose your shirt or go home with empty wallet.

Both:
One is that past results matters.
One is because they tend to repeat by nature.

Reason:
One reason is hot numbers or bias numbers or the tendency of fluctuation favoring certain numbers.
Numbers seems to come in clustering periodicity - waves - and the optimal window is 5 or 6 ...

A famous bias player claim that you can reduce the house edge playing the last 5.
He also claim that past numbers do matters.

One thing i can't get out of my head is that if you pick the last 6 numbers with no repeat, then you get one repeat with the next 6 numbers.
Also work with 5 numbers and the next 5 number window.

You track the numbers on a rolling basis where each window is 5 or 6 numbers.
+ sign is when you hit a repeat within next window of numbers.
- sign is when you miss a repeat within the next window of numbers.

+ + - + - + - + + + + - + - + - + + - + + + - - + - + - + + + - + - + +

Is there any way to take advantage out of this clustering appearances of repeats.

congratulations [smiley]aes/wink.png[/smiley]

Buffster i test this the reverse way.
I intended to flat betting and win +1 or break even +0.
If i did not succeed i let the fibo progression show me what kind of loss i would end up with - showing me the negative expectation or variance.
1 1 1 2 3 5 8 13 21 34 55 89

I run 40 sessions and end up at most with -2 units with highest bet of 3 units.
Then i got some kind of variance indication of the threshold.

At sport betting forums you can read that you should never risk more then 5% of your bankroll.
Sound easy, but is pretty hard to achieve.

 
Strike domination from the very first outcome to continue or else i aim to break even.
This is the results for 10 sessions.
This is using a kind of regression strategy.
Won 59.5 units and lose 30 units.
[reveal]

-
21.0
-3.0
-
8.0
-3.0
-
4.0
-3.0
-
0.5
-3.0
-
8.5
-3.0
-
5.5
-3.0
-
2.0
-3.0
-

0.0
-3.0
-
3.0
-3.0
-
7.0
-3.0
-

I like this kind of method, as domination or strikes compensate for the negative expectation.
And i also like that you get a second chance to break even before you hit the hole.

+0.5
+0.5
+2.0
+0.0
+0.5
+2.0
+0.5
+0.0
+0.0
+0.0
+0.5
+6.5
+0.5
+0.5
+3.5
+0.5
+2.0
+0.5
+0.5
-3.0
- - -
+0.0
+0.0
+2.0
+0.5
+0.5
+0.5
+2.0
+0.0
+0.5
+2.0
+0.0
+0.0
-3.0
- - -
+2.0
+2.0
+0.0
-3.0
- - -
+0.5
+0.0
-3.0
- - -
+3.5
+2.0
+0.0
+2.0
+0.5
+0.0
+0.0
+0.5
-3.0
- - -
+0.0
+3.5
+0.5
+0.5
+0.5
+0.5
-3.0
- - -
+0.5
+0.5
+0.5
+0.5
+0.0
-3.0
- - -
-3.0
- - -
+2.0
+0.5
+0.5
+0.0
-3.0
- - -
+0.0
+0.5
+2.0
+0.0
+0.5
+0.0
+0.0
+0.5
+0.5
+0.5
+0.0
+0.0
+0.5
+0.5
+2.0
-3.0
[/reveal]