Collecting roulette ball jumps information


highly recommended

Thank you for suggesting ways , always effective impact !

all the best


I am analysing the scatter and the ball jump scatter varies a lot. The wheel at my casino has a bias for a dominant diamond. How to take advantage of it?

What is Laruance Scott RA?


Otter how do you make scatter profile

For example are you taking rotor speed and write it down for that particular trail and then watch wish deflector ball hit and write down number below deflector and the final number where ball stop

I think that some one should look for overlaps when collecting scatter profile where one particular deflector will donate ball jumps into the same area as other vertical deflector hits.

I assume there exist more easy and more advance ways to create a scatter profile and there is different ways that is better depending on if some one use a Roulette Computer or VB

RA stands for Roulette Analyser and you need to read Laurance Scott volume 2 to fully understand and operate program - but classical ways charting with pen and notepad works great - but software is more advance




This is the ball jump I collected. I record jump from the number the ball first drops on to the final stop.

Here is my data. I am not sure how to create the graph and to get the average. Some help would be nice :slight_smile:

I collected about 105 spins

Jump Count
0 2
1 6
2 3
3 5
5 3
6 3
8 6
9 6
10 1
11 7
12 3
13 6
14 3
15 4
16 3
17 3
18 3
19 6
20 3
21 2
22 2
23 5
24 2
25 2
26 1
28 1
29 3
30 2
31 1
33 1
34 1
35 3
36 1
37 2

What I observed was that the scatter depended on where the ball hit the diamond and the rotor deflectors. Rotor speed matters if it is very fast. With fast rotor the ball most of the time bounce back

Here is a graph I tried to create on excel

How do you select what ball jump to play?


Double zero wheel ?
Best place is 12 , but i very doubt if possible beet such wheel, maybe with some more information. Here is the case when radial counting not do good answer…


It is single zero wheel.

I can collect more data. but from the graph I can see there 2 major peaks.

Is it good to bet on both peaks?


Here is only one pick all other what you see in your chart is simply dispersion.
More data if to look as to jumps not change nothing essential . You must collect different jumps from different diamonds , but i doubt if in radial way that still get something. Maybe if you will find some relationship with wheel speed…
I always collect data in linear way and of course very look to relationship with other paramethers, but that not mean that such wheel will be playable in mine opinion. I think that is wheel without selection and that in mine opinion is not good…



How the ball could jump minus 15 times to position 31.

I believe you mixed standard deviation graph with jumps.

To have 10-SD it must be significant amount of data and often surrounding pockets get added and averaged, for that reason the graph looks as you showed.


In mine chart are not STD. I said that such wheel maybe not playable…
And meaning -15 is not jump back . Not pay atention what is on chart coordinates.




This is same your data.
First i moved 2 hits from 37 position to 0,
If you start with 0 you can’t have 37.

Your 105 spins i grouped in groups of 7, (can use 3 or 5 or more).
For example position 10 gets added, 7,8,9,10,11,12,13 and divided by7. (don’t have to divide)
That gives me smoother look.

Orange line is an average of 105 spins.
Could divide with 36 instead 37 to overpass casinos advantage.

Even position 10 may look the highest, it is to close to the edge , so might be better to play around 15 in the center.

Of predicting spins with roulette computer the ball may not in average drop exactly on predicted number but it may drop say about 3 pockets earlier. In such case you would adjust offset to 15 + 3 = 18.



Are you saying exes on the chart are wrong?


No that is simply amounts of hits 4-x-4 . I always do such charts for mine play 9 is best amount for me so acordingly such calculations mine results are best…


How amount of hits can be -15?


that is not -15 that i do simply because will be better to see which zone is positive which not .
All what with minus are under average - all what in plus are above average…


Here is chart in STD. Really almoust random result.

If we cover very precize what is on 18 pockets we can have 3.5 STD, but that is not realistic disperssion +/- 3 pockets from center line gave the same near to 1STD , what is not enough for winings in mine opinion…


At 18 pocket i see only 0.7 value, so what now have to add to each value 2.7.


I have in mind if to cover 18 pockets - so half wheel . Upper chart is for 9 pockets . If we will cover idealy precize so say we look to data and see what we can best to bet, so in past data we can find place with 3.2 STD. But if look to front then our STD is only 1.

Bigest mistake do who count STD to back side and find 3 STD thinks that such is good for play. STD we must count always in front so firstly select what we bet and then calculate how much times we will hit.
When i collect data i always mark all hits in few groups in about 15-20 spins, this way i can calculate which is STD if i play acordingly results of previous group… that much exact show what we can expect in real play.

I separate collect jumps only for research purposes. For using i collect all distances from some for me stable point.


But that is not SD for 18 pockets.


STD is: SQRT(NPQ) where N=103, P is probability of winning and Q is probability of loosing.

I not know now exact name of this value in english must look , but the same calculate many players . STD is counted in sigmas and show probability that result is not random…

3STD ( sigmas ) in front is 99.9% that result is not random , that are some reason for ball stoping or falling here , 1 STD ( sigma ) chance is about 66% that result is not random…

You can count for himself what you want from chart to back side …To front we can only estimate, so calculations which are in chart are acordingly mine estimation…

So what will be usually , if i will have such results to back . Because results to back itself say nothing usually.
Must be - prediction / result this way we can calculate STD for prediction point.


Thanks for the explanation but you always explain something else and not what you quote.

Perhaps you could say yes or no.

But that is not SD for 18 pockets.

You write explanation what the SD is.

I am not an expert in data and probability but it is logical to me that if you look SD across 18 pocket SD value on same amount of spins would increase, it may look as you have less SD but in reality it may not be the truth, because and the mean increases more same as the amount of hits from the mean.