Simple VB Method + Excel (Scattering Analyzer V1)

Hello everyone, thanks in advance to those who read. I did a research on this forum and it seems to me that many here can comment on an idea.

I practiced around 150 hours with different visual methods, all on live online roulette. I mostly do the sessions with an Excel spreadsheet where I measure the position differences between two temporal points. The final point is always the result, and the starting point is varied according to “the method”.

The one that convinces me the most so far is the following:

Suppose I can determine when one revolution is in 1 second. As soon as this condition is met, I observe the number under a diamond (always the same) and load that number in Excel. Then I write down the number that came up.
That pair of numbers (initial and final) has a certain separation, which is graphed. A point is not plotted, but with a margin of ±3 positions relative to it, since those will be bet. The width would be 7 consecutive numbers in that case.

After a certain number of shots I can get something like this, with three different widths (number of numbers played: 3, 6, 9):


Roughly, if a triangle is generated, the center point will be the midpoint of the bet.


First, I get a delta of 15 and 0 (end - start). In the case of 0, it means that I read the same number that came out.


For this reason, I am looking to bet on the maximum of the curve obviously.
End example.

This can be seen as a scatter analyzer depending on which method you use. One could be VB2 obviously, I recommend trying it. Lately I proceed as follows:

  • Determine an average rotor speed according to the dealer. Record the first 6 Δs and average the measured rotation time for each shot. Suppose it gives 6.5 sec.
  • I observe at what speed the rotor starts. If it is between 6 and 7 seconds, I proceed (possible variation of this rule).
  • When spinning starts, I have an imaginary click every time the ball passes through a certain diamond, any, but always the same.
  • When I detect that approximately one second passes between clicks, I read the number in the chosen diamond. This is because according to my experience, that is to say the roulette wheels that I play, when this condition occurs, around 11 ± 0.9 seconds pass until the ball begins to fall. Of course somehow there must be a certain trend; not necessarily a dominant diamond.
  • I write down that initial number in Excel and it generates the curve.

The Excel has 5 columns that are used to simulate bets and also make them online:

1: Signature: this is because I started considering the number that came out as the initial number (wrong). I note according to the chosen width the central (or maximum) number of the curve, if that triangular shape is generated.
2 to 5 are automatically generated.

2: BET: The position to bet on automatically appears there. Numbered from 0 to 37, starting from 0 in a clockwise direction.
3: RESULT: Result of the bet (1 if you win, -1 if you don’t).
4: ACUM: measured in chips, but partial accumulator according to the bet width.
5: ACCUM. Accumulated sum starting at 0 (positive → amount of chips won).

I hope you can try it. The results so far are positive with about a 10% win on 412 total shots, 157 played, 286 chips won, with a margin of 286/(157*7)=26%.

I would like to know how it works for others. For now we are positive, with enthusiasm and anxiety. I’ll keep you posted if the method still stands or not. Suggestions and tests would already help.


  • Scattering is highly dependent on rotor speed. I suggest setting an average speed, and playing 40% of the shots or less perhaps.
  • Boxes can be added or subtracted quickly if the initially estimated rotor speed varies a bit. If the average is 7.7 cells per second, I add or subtract around 3 or 4 on some occasions. Do not get to 7 because that implies a variation of 1 sec in the rotor, enough to lose the triangular shape.
  • The BI column is automatically generated and corresponds to the plot. You can delete values ​​if you prefer not to plot a certain shot. For example, those that end up spinning indefinitely in the rotor after having passed the diamonds.
  • If the spin alternates direction, simply use two sheets.
  • You can try to vary the observation time to less than one second. It would be nice to get it, although I understand precision in how much the rotor turns more or less is resigned to an error in the reading (by measuring before, the more time the rotor spends turning). I think the key is to estimate as accurately as possible the initial velocity of the rotor. I find in my case that counting down to plus or minus a second is not as difficult as quickly estimating how fast the rotor is turning.

I hope it is useful, I would like to know if it helped anyone.

Best regards from Argentina

Scattering Analyzer V1.xlsx (244,9 KB)

You know - all that you wrote is quite hardly understandable, even for a person who more or less can code…
Are you that Ignacio who in other forums creates systems? :slight_smile:
If the answer is yes then all is more or less clear, because from what you write is clear, that you are not a VB player…

For example, what you wrote - " one revolution is in 1 second" Let’s say you somehow can recognize that. But that is simply not enough - at a minimum, you must know where is a ball in that moment but read the number not under a diamond, but under a ball.
And even if you are able to do that so know where the ball reaches such speed and read the number under the ball still such a method hardly will give you a big advantage, because the same speed - does not guarantee that the left distance will be the same…

The aim is to know the left distance…
If you will find say some method on how to measure ball deceleration - that will be really worth something…
Now what you write is nothing new and worst of all this not gave such an advantage about what you write…
No need to write in the forum about a +26% advantage - much better simply to go to the casino and take money…


You have to have someone from inside to win :rofl:

The example was incomplete. Sorry!


In the first two throws I get a delta of 15 and 0. In the case of 0, it means that I read the same number that came up.


I add one more shot (third), which generates a peak at 17 (Δ) because if I had bet on said increase in these three shots, he would have won two. If you had bet on Δ=2 or 35 you would have won once. At 28, he would have won 0 times. That is, the vertical axis indicates the number of hits (or hypothetical hits) as the game progresses.


For this reason, I am looking to bet on the maximum of the curve obviously.

End example.

Jajaja no no I’m new here, not that Ignacio. This is my first roulette “software”.

Excel is the best value of this publication, being able to use any number as a predicted number according to the preferred method. It is to analyze rather the dispersion of the prediction, beyond a simple offset, of course.

Regarding the method I used, I look at a diamond and I count when the ball takes one second to go through it. The count starts at the diamond, and ends at the diamond, so I see the number under the ball and also under the diamond. It doesn’t have to be exactly one second. I noticed that under this condition, the remaining time (to the fall) is constant with a relatively low dispersion, and if the wheel is titled of course, some area will be predominant (and this is seen as a triangle in the graph). Sure I’m not inventing anything new, haha, but combined with Excel I think its usefulness can be better evaluated.

So we all combine with Excel and not only with excel…

here better for you simply to do some examples, because when the file is empty is very hard to understand how you want it to use…

If you are from Argentina - you must know Toby … ?

Yes, that was Ignatus, but who knows - can change a few letters :slight_smile:

I do not think that I am last in the Vb, but I absolutely do not understand about what you here … ?
Maybe somebody understands?

And why is bad simply after detecting ball speed - to mark the starting point and calculate the ball distance and do some chart say like this
I think from this chart all is clear where we need to play and how big advantage we have on Y axis is STD

What about selecting a wheel with dominant diamonds hit?
Adjustment for rotor speed change?

Hi bebediktus, forester and everyone.

Let’s make an example from scratch. It is essential to follow it together with the empty Excel spreadsheet. These numbers were taken from a real session. They are related to two consecutive ones (the first is the observed number, the second the result). The observation, remember, can be taken with any technique. In the following case, the initial number corresponds to the moment in which the ball takes one second to make a complete turn, taking a particular diamond as a reference.

Starting in Excel cell A8, let’s copy the first 8 shots. There are 16 numbers because each shot has an initial point (observed) and a result (where the ball actually lands).


You should have reached row 23 when entering the number 6.

Note: the three curves correspond to 3 different “widths” (explained in the first post). The gray one is the one I usually use, which includes a total of 7 numbers (3x2+1). The red one has 13 numbers (6x2+1) and the blue one has 19 numbers (9x2+1).

Note that in the gray curve generated, there is a clear trend between approximately 16 and 25. There is not a maximum but 2, in 18 and 22, so we can use a midpoint as the best estimator (20).

This means that, for example, if the reading (prediction) corresponds to 0, the number that will tend to appear is located 20 cells later, clockwise, that is, 24 black.

Suppose we decide to enter to bet, or simulate bets from now on.

In cell BM25 write the number 20. This is not the black 20, but the expected increment (called Signature; it’s not a Dealers Signature, it is just the expected increment for the technique used).
I pass the rest of the numbers to copy, but don’t copy all of them for now.


We add by two, that is, we are evaluating what happens in one shot. We copy the 5 and 19. As soon as we copy the 5, since we already introduced the 20 in Signature before, the 2 will appear in BS25. This means that we must bet on the number that is in position 2 (15 black) and its neighbors (3 on one side and 3 on the other). The 19 came out, which is in position 3, and we won. In the BI column the resulting increase between these two numbers (5 and 19) appears, and it is indicated that it is 21 positions for this case. Now we are 29 chips up.

If we are doing well up to here, moving forward I think will be easy. We are going to continue adding two numbers at a time, and if necessary we will change the Signature based on what we are seeing on the curve. We keep the 20 for now and write it to cell BM27.

We copy the 26 in A26 and we see that, when doing so, in BS27 the 19 appears (remember it is an absolute position). We copy the result, 6, and see that the prediction was not close. 7 chips were lost.

Keep copying two numbers, that is, one shot, and choose the Signature according to how the curve changes. Attached capture of what I chose for this case.

I hope to hear news about whether you find it useful. Let me know if you need help with anything…

Also keep in mind that this chart can be done on paper by playing live. With some simple grids I estimate it is possible.

Thank you group, best regards.

Yes Sir, it is ideal to find that situation. In my experience I have noticed that many times there is a certain trend but in a contiguous area. For example, in a whole third, or half of the wheel…

Regarding the adjustment for different rotor speeds, I estimate that the change in speed does not imply an offset that is easy to calculate. I mean, if the new speed generates that at the moment of the fall, the rotor is 10 positions more advanced, it is not correct (I think) simply to add 10 boxes to the prediction. This is because the ball will land 10 positions ahead, but it will continue to bounce, and that bounce is different if the rotor angular velocity changes too much.

A clear example would be, what happens if the rotor changes speed so that it advances a complete revolution? Wouldn’t offset be necessary? I think that it’s not like that (some kind of non linearity), because the ball will initially hit the same number, but from then on it will have a very different behavior depending on the speed of the rotor.

I prefer to stick to one particular rotor velocity with ±0,5 seg of variation. I’m not sure if this is correct, what I’m saying is pretty empirical, and I’m probably lacking some experience.

For now, all is clear what you calculate. Initially, I not noticed that you hid some columns. Why not use a simple option - to hide? And at all much more simple to write a few small macros…
Ok done as done. I also use something similar to simulate my expectable results, only I do that much more simple and more flexible - Widnes I can choose free which I want.

But the biggest drawback is that you don’t take rotor speeds into account. I understand that you’re playing within some limits, but measuring the rotor and calculating its path to the end is probably the basis from which the game itself starts.
Of course, you may not calculate that path, but just link the rotor variable to the result and calculate the correlation function. This could give better results, but it is much harder and requires more experience in gameplay and probably more programming experience.

You have to realize that the more you leave room for randomness, the worse your results will be.

There are 5 main variables in this game that affect the outcome. These are ball and rotor speeds, ball and rotor deceleration, and scattering characteristics. All of those variables affect differently depending on some other variables, even ones we don’t think about…

It is very difficult to understand to what extent the influence of each variable should be taken into account and a program that can solve this problem, most likely, has not been created yet…

But I believe that something can be created because I saw a game of player who plays cardinally better …
So that is my engine, I want to discover something that I could play similar to him… :slight_smile:

What you mean by “±0,5 seg of variation” ?
Rotor speed not depends on you it is as it is, so what you prefer is in last raw, you can choose to play only some specifically speed, but my observations never show that this leads to significantly better results.

I couldn’t follow here where it was written there were no 23 and 6.

4 and 4.5s rotor may not have different ball jumps but it may give you 10 pockets difference.

Close to 1 rotation rotor difference is a different story, you tune your prediction to such conditions and use it when the rotor is in that range.


Ok, but there was no such picture before.

So it I did when putting the numbers into his file.
I use something similar when calculating what can be from my data, only he did the file in quite a strange way - simply things did in a very difficult way :slight_smile:
In simple words, he marks some Reference points ( at 12 o’clock when the ball is the first time slower than 1000ms) and the second number is - the winning number. Instead of writing them into two columns separate reference points and winning numbers - he writes all in one :slight_smile:
Then he finds the distance between that Ref point and the winning number which repeats most often and writes it to another column and that means that here he does bet. And in the last part of the file, he calculates how much he will win or lose if he will play this way…
If he still will include the rotor way, will be something very near to what we do …

When I hide the columns, I don’t know why errors are generated in the calculations. I don’t know about macros yet but I guess at some point I’ll learn.

Ok, fine. I think it can be done in a way that you have probably already used, but considering the following.

It seems to me that dealers do not spin the rotor with random speeds. I find that most, after looking at about 10 shots, can have a relatively normal distribution (with some outlines). What I try to do is determine said normal or average speed, and play only those that are in a certain range with respect to that average; ±0.5 seconds I think it allows to enter a lot. When starting the shot, observing how much the rotor advances, for example, in 2 seconds, I estimate that it can give an idea of ​​whether it is faster or slower than said average and make corrections quickly.

Also I try not to play in the range of 4 seconds per revolution. I think it is too hard, and maybe 6 sec/rev are not uncommon.

Thank you for reading and you answers untill now!

Thank you forester for the observation, you are right of course.
I can’t handle yet such speeds, 4 sec/rev it’s very hard for me. If we can find 6 sec/rev aprox speeds, the difference within 11 seconds may be around 5 pockets. We can consider them as usually when rotor starts moving, no matter if it’s 4 or 6 sec/rev, but the Excel doesn’t have an option to do it. We can add those pockets manually in different ways I think.

1 Like

Anyway even if they want to push the rotor with ideally the same speed they do mistakes always and that mistake can cost you much say two pockets in 2-3 sec so you can look at the same speed but finally, that creates 10-15 pockets difference, what absolutely change win/loss situation…

so imagine that you clock rotor speed and know which mistakes you can do, so wheel speed you write as data and add possible delta for error.