Evolution of roulettes advantage play

Where was the start and where will be the end?

What options has an advantage player to win at roulette?

Let’s look at roulette game.

Dealer spins wheel in one direction and ball in the other, after some spinning around the ball hits one of diamonds then drops on rotor where it continues jumping until it rests in one of 37 pockets.
There are many ball rotations sometimes more sometimes less, many diamonds and many pockets with numbers. Even rotor sometimes is going faster. We assume at each spin different amount of rotations can be made, any diamond may get hit and ball may jump every time differently.

It looks random, and it is if you do not know anything else and do not look for more.
But what is in there to know?

Well let’s go step by step and if you finish reading all steps you will start understanding, if you already understand you may understand it better or perhaps help me, or other readers to learn more.

Next to each roulette table you can see display board where it shows past results, some boards may show the most hot/cold numbers in past 300 spins. Roulette players love to use data from the board, mostly for all kind of stupid reasons but here I will focus on one which relates to what we call “dealers signature”.

The display board may display ~20 past results it may look as this;
22,6,17,22,32,23,0…etc. where number 0 is the last winning number at the table

Sometimes I observe players to see how they place their bets. I found out that many of them believe they are playing dealers signature and that it may help them to win. Reality is, 90% of them are not doing it properly and with full understanding.

Dealer’s signature (DS) from display board brief explanation

There is a belief that some dealers may throw ball with same strength, spin rotor with same speed and that ball may jump similarly therefore we should be able to compare spins and find out what outcome may be.

If we look the board and if zero is last winning number;



Because the ball with each spin alternates direction (clock way or anti clock way) we have to look every second spin.
For example if based on DS we want to know what most likely winning number will be after last spin win at “0” we need to look pocket distances in between 32-23, 17-22 and 22-6.

From number 32 dealer spins number 23 which is distanced +16 pockets
From number 17 dealer spins number 22 which is -17 pockets distanced
From number 22 dealer spins number 6 which is -18 pockets distanced

Obviously -17 and -18 are close and we can say for 2 spins ball stopped on opposite side.

  • 16 may look as opposite but in reality is not since wheel is a circular so it is only 3 pockets from -18.
    +16 is same as -21.

After understanding it, we can say for all 3 spins in particular ball direction the dealer hits opposite side from number where he picked the ball.
If we want to play this kind of DS after “0” we would need to place bets on opposite numbers from zero.
Such as 5,10,23,24,6,33,8…

If we look opposite ball direction we have ;

If you analyse this same way as we did previously, you will find out that there is no DS for this ball direction.

This is only a small sample, player may follow spins for longer but I’ll need to disappoint you, if that’s all you do to play DS it will not work.

When I said 90 % of people do it wrongly I mean it. Many people will not even analyse data for different ball directions separately, so they look distances regardless in which direction ball was spinning, of course that’s very wrong.

Some players may say, I know this dealer after number 22 usually spins 17 or 33, hmm 17 or 33. Really, 17 and 33 are two different worlds; obviously their opinion is made based on coincidences they noticed.

Let’s continue.
Back to our displayed results.



Yes the ball stopped on each of displayed numbers, let’s assume dealer each time spins 100% same spin.
But there is time delay in between how long takes dealer from spinning the rotor until he spins the ball.

For example he may pick the ball from zero, spin rotor in opposite direction and most commonly spin the ball after 1 sec. But what happens if he spins the ball after 2 sec? Rotor already moved by 10 or 12 pockets extra. Therefore instead of following numbers from display board we can observe from which number dealer spins the ball. It will increase accuracy of DS we play.
Also it lets us see how spin looks at start, if it is very different rotor speed or ball speed at start, we have option to not place our bets. Many people even seeing different spin if result match their expectation still believe it is result of dealers signature. No it is not, it is a coincidence!

Yes there is more, and everything explained until now still is not good enough to make profit at roulette.

Who discovered a “tilted wheel”?

I would nominate him for the Nobel’s Prize.

Roulette wheel is sensitive device, smallest angle on table may cause wheel to tilt.
Only 1% tilt may cause ball to exit ball track at one particular place 70% of time.
For analyses we can look at it as 100% (dominant diamond hits).

Let’s say each spin we know at which diamond the ball will hit (dominant diamond DD). Imagine rotor is not moving, at 12 o clock diamond we have number zero. Dealer spins strong spin, ball is spinning and on the end it hits 12 o clock diamond, then drops at number zero. Next, the dealer makes spin shorter, but same happens the ball again hits 12 o clock diamonds, and then drops at zero. No matter how strong he spins the ball we get ball to drop at same place.

Let’s look what will happen when the wheel moves 1 pocket per second.
Dealer spins the ball which will spin for 20 sec, for that time rotor moves and because it’s so slow it makes only 20 pockets in 20 seconds. If dealer start spinning ball and under DD is number 0, at the end of spin when the ball hits DD under DD we will have number 24, since number 24 is 20 pockets distanced from zero (CW clock way ball direction).

As you can see the ball because of tilt couldn’t drop at 9 , 6 or 3 o clock diamonds which could produce any results. Now our distance in pockets “d” from start of spin until ball drops depends on time “t” that ball spins during the spin, and rotor speed “v” in pockets per second.

d= t x v

For example 20 sec long spin and 10 pockets/sec rotor speed will make rotor move
by 20 x 10=200 pockets.

If we divide it by 37 (amount of pockets) we get that rotor will move by 5.4 rotations, or 5 rotations and 15 pockets. If at start of spin under DD we have had number 0 when ball hits DD we would have number 30, which is 15 pockets distanced from number zero.

What if next spin dealer spins, same rotor speed but the spin is 20.5 sec long?

It is very important here to understand that if wheel is not tilted that extra 0.5 sec may be more than one ball rotation, maybe more than 37 pockets change, and that’s why all previously explained about DS would never work.

But on a tilted wheel it is
20.5 x 10 =205

Wooo, only 5 pockets difference compare it to more than 37 pockets if the wheel was levelled.

With levelled wheel we look all number pockets that ball crosses, with tilted wheel we look only pockets that rotor moves.

But there is more, even this will not work and produce you enough advantage to win at casino.

The wheel will never be so tilted that it hits only one diamond, ball spinning will never be constant, it may be 2-3 sec different which even with reasonably slow rotor speed as in our example makes 30 pockets differences also rotor speed is not constant.

15 or 20 sec long spin is relatively long if rotor is 1 pocket per second faster it may result in 15-20 pockets error. Ah, and that ball jumps can’t be forgotten. When we add it all up we get nothing?

Sure this may be simulated at home and proven to work with an old wheel, controlled environment, shorter spins, slower rotor etc, but in todays casino it will not work.
Twenty years ago when casinos did not know about tilted wheel affect (many wheels were heavily tilted), when wheels have had deep pockets, when balls were not jumping so much DS did work.

Once when you finish reading this thread and see player’s chances when he uses more advanced techniques you will understand why all explained until this point has no or it has only small theoretical chance to work.

Some time ago I made a post
Roulette, dealer signature myth or not

Also it is good to read
Genuine winner or similar roulette systems

Try to understand chart displayed in that article.


Reasonable roulette spins are about 15 -20 sec long.

It is too long to expect that from spin to spin everything will be same or close to same so we can expect similar results.

Same rotor speed
Same spin length
Same deformations on ball track
Same ball jump
Same vibrations
All the time hit at one DD

Take some time in casino at one table, observe steady dealer, all you need to observe is how many rotations rotor will make from start of each spin until the ball drops, and you will understand it all.

So, the advantage play developed further. Why not if we can? ;D

Simple rotor timing can tell us if rotor speed is really same as with previous spins, is it faster or slower.
If we detect faster rotor even if we add only one pocket to our prediction we will be more accurate, but definitely we can do better than that.

If we reduce spin length by predicting it lets say 5 rotations (or 5 sec) after it started we are again more accurate it’s because if we have some errors in rotor timing it will be less since remaining time is shorter, if ball track effecting ball differently from spin to spin that will be reduced as well.

But if we read number under DD five rotations later from start of spin it still doesn’t help us to get constant time that ball will travel until it drops. One spin may have 20 ball rotations, another one 17, it still means we will be reading number under DD once at 15 and next time at 12 rotations before then the ball drops.

This may look as a solution but it makes it worst, for us consistancy in ball remaining time is important, 3 rotations difference when ball is faster (20-17) are shorter in time compared to 3 rotations when ball is slower (15-12).

(Until ball makes 3 rotations from 20-17 if average ball rotation is 0.4 sec all together is 1.2s
If we look 15 to 12, ball rotations are slower, let’s say in average 0.6 sec per rotation. 3 x 0.6=1.8 sec
1.8s is 0.6s longer than 1.2 s, that’s’ why it is worst. This is important to understand.)

Therefore to reduce length of spin we need something better then just counting and deducting fixed amount of ball rotations from start of the spin. Something that can identify particular ball rotation so in every spin we can have remaining time of ball traveling constant as much as it is possible.

If you have some knowledge and follow this thread you could notice how we moving towards visual prediction methods.

Traditional roulette visual prediction

From everything previously explained related to tilted wheel we know that if we enter every spin when remaining time of ball traveling is ~constant, the result where the ball will drop depends only on rotor speed , or we can say rotor movement until ball makes remaining rotations. If we still consider for every spin that the ball will hit same diamond (DD) it also means that after ball is passing our DD at our particular time during the spin the ball will make fixed amount of whole rotations until it drops.

We already learned why we can’t deduct ball rotations from start of the spin. It doesn’t help us to have same amount of ball rotations from end of the spin.
If for example we want to predict in 6th rotation from the end when remaining time of ball traveling is about 10 sec, if the spin is 20 rotations long we can’t just deduct first 14 rotations.
If next spin is 17 rotations long and we deduct 14 ball rotations, we will have prediction in third rotation which will give us completely different result not related to previous one.

Therefore we need to use something else.

All visual ballistic concept is based on how to identify particular ball rotation from the end of spin, so after that rotation we can have constant ball traveling time until it drops.
For example if we want to detect when during the spin ball is 6 rotations to the end how can we do it?

The simplest way is to observe spins and to estimate it. After some practising we will be close with results and estimation. We can stand next to the wheel observe it then in one moment read number at our DD, then count remaining rotations that ball makes. If after each attempt we get six remaining rotations we do it well. But can we really do it on so easy way. The answer is hardly. Sometimes after estimation the ball will make 5 rotations sometimes 6 sometimes 7, most likely we will be in that range.

This is what Jafco encourages in his ebook, advertising it as a revolutionary way of prediction.

He made nice videos for training where on his wheel is possible reasonably well to guess particular ball rotation 3 or 4 rotations before then the ball drops. At that wheel ball slows down extremely fast about 300-400ms per rotation therefore it is easier. On real casino wheels it is more 130-200ms per rotation which makes it harder. To highlight it, ~200ms difference is the best that we would get at the best time moment during the spin, predicting 3-4 rotations before ball drop may be only 100ms difference which on 1.6s long ball rotation is not much. For example it’s much easier to spot difference in between 1000 and 1350 ms per rotation, compared to 1000 and 1150ms on wheels in good condition.

Personally I believe 3-4 rotations before then ball drops is too late , just estimating is not good enough also there is a bigger reason which will be explained soon. Considering that he uses electronics timer to clock rotor but doesn’t use it to find particular ball rotation really is not justified. However most people who bought his system with previous knowledge close to zero are happy, everyone has to start somewhere.

There is some discussion about Jafco’s system here

Jafcos ebook explains reasonably well principles of tilted wheel prediction but it doesn’t have solution how to define particular ball rotation during the spin, except estimating it.

Yesterday I looked his ebook again just to make sure i do nto make mistake.

I was searching if there is anything explained about ball deceleration knee point during ball spinning.

I did not find anything so I will explain it here because it is important for a VB player who by estimation is targeting particular ball revolution.

Better way of estimation is if we target particular ball revolution where time difference in between two rotations is greatest.
It’s easier to detect it, since at that moment ball rapidly deccelerates.

In roulette world this is called the knee point. The knee point on some wheels may be strong, it can happened in two ways that in particular ball rotation the ball decelerates rapidly, or perhaps right after the ball deceleration rapidly drops. It is never 100 % accurate but in some cases it may be accurate enough to produce an advantage and definitely it is a better way to predict. Such point in usual is when the ball is about 1-1.3 seconds per rotation (or 5-6 rotations before then the ball drops), that gives to the player possibility to predict earlier. When prediction is earlier even if particular ball rotation is missed error created is smaller since time of rotation is shorter. The knee point can be noticed when player looks for sudden change in ball speed, if he can hear the ball sound it also may help.

Picture above shows times of ball rotations.
Data is taken from roulette computer, after 17 ball rotations were clocked and processed.
As you can see on the chart first ball rotation is about 500ms then with each next rotation time increases. Important here is to notice point marked with red arrow. You can see that next rotation after that point decelerated more, therefore time of that rotation is increased by more than what was happening with previous rotations.

Also notice how deceleration is dropping 3-4 rotations before the end. That is why on most casinos wheel is not recommended to predict 3-4 rotations before ball drops. At that stage of roulette spin ball is leaving vertical edge of the wheel, it may even drop by let’s say 1 mm, friction is smaller but and radius is slightly decreased, drop in ball deceleration is much higher then you may expect or as an average Joe can spot from this graph.

This orange graph shows how ball decelerates per ball rotation.
Differences in milliseconds of one ball rotation compared to previous one.

Red arrows clearly can show the knee point, you can see that previously differences in between ball rotations were 50 to 100 ms and then suddenly we get 250ms.

Finally this green graph shows real ball deceleration not per rotation but per second.
We know that rotations on the end of spins are getting longer so we are comparing longer times therefore previous graphs will show greater differences.

This graph is real representation what is happening with ball in time.
We can more clearly see the knee point but more importantly as red arrow is showing, we can see how ball deceleration is rapidly dropping towards the end of spin. This explains why predictions by estimating particular ball rotation shouldn’t be done 3-4 rotations before ball drops on casino wheels.

And again same question, can someone understanding all of this finally make any profit in casino?

I would say yes.

Now, the time from prediction to the end of spin is not 15-20 sec long that can deviate but shorter and reasonably constant.
Until now the player knows that he needs to time the rotor and make appropriate adjustments, if he is skilled to estimate let’s say 6th ball rotation most of the time, if the wheel has reasonable tilt, if ball jumps distribution is not close to random sure he can. Better conditions better his chances are.
But there is more…

Scott observes ball crossing rotor numbers and for example when ball crosses 0, then double zero (American wheel 00 is on opposite side), then zero again, if rotor is let’s say 3 sec/r he knows that ball is about 1.5 sec. So when he finds such crossing pattern he reads number under reference point and relates it to ball drop point. If you practise, scanning for cross pattern is reasonably easy. Problem is that ball never makes exact crossing and that we still do not know at which point is braking point in between 2 rotations but that is problem with any VB method. The ball can be 1500ms per rotation or 1550, with cross pattern we can see it same, but 1550ms ball can actually make one less rotation. (This same applies if we are just guessing for particular ball rotation.)

Since Cross patterns are using rotor as a reference, in this example rotor speed was constant 3 sec.
If rotor speed changes only slightly to 3.1s now we find perfect cross over pattern and real ball speed would be 1550 ms (3.1s / 2), 50ms slower ball can make one less rotation to the end of spin, therefore rotor traveling time will be shorter and reference number will not match previous predictions.

We also can analyse what will happen if rotor speed is 2.8 sec/rotation. We will get pattern when the ball is 1.4 s/r we will read our reference number earlier so for example instead of number 0 we may read number 14 (European wheel ball is in clock way direction), but in reality that ball will make one extra rotation, so it will drop at zero + [rotor change to the end of spin ~13 pockets]. Slight change in rotor speed makes us double error (instead of 0 we predict 9 but the ball drops around number 13. Sure the player will make adjustment for rotor change and eliminate 13 pockets error but there is still error from 31 to zero.

Imagine if these two differences can be reversed to each other, it would be perfect way to play and in most situations we wouldn’t need to have additional adjustments for rotor speed changes.

All this means that the player must have perfect feeling for rotor speeds to be able to play cross patterns. Many times it’s been said, one in 1000 will be able to do it.
If you followed you could notice that predictions are late, 3-4 rotations before then ball drops. That is additional problem. On most wheels deceleration of ball at that speed ranges drops, which makes it harder to differentiate in between 2 ball rotations, but also if we miss particular rotation, we have remaining traveling time for rotor wrong by about 1.5 seconds.
If rotor speed id 12 pockets per sec that will produce error of 12p/s x 1.5s= 18 pockets.
But this was example with 2 x cross pattern surely we can use 3 or 4 times, or any variations.
Some people are so skilled they read numbers when ball passing and instantly know pocket distances in between them. There are similar methods to Laurance’s cross pattern, all similar ways but and similar problems.

Check UWE

Cross patterns is very old method to play, just his idea of taking advantage when roulette wheel is tilted is great. Laurance introduced and different approaches to visually predict roulette. Another way of estimating particular ball revolution is by listening sound that ball makes and looking for the knee point (sweet spot) even using a timer. You still can buy his material onlin. Be sure he will not promise you easy winnings (as some) and tell you everything you want to hear just to take money from you. Scott is a knowledgeable person and experienced player, definitely not just a salesman.

Roulette players will never give you a guaranty that you will win, because they know how much it takes to be a winner.

But there is more…

Let’s introduce some technology to roulette advantage play.
Do you remember disco strobe lights, one that starts brightly flashing in about 1 sec intervals?
With strobe light on, everything around us slows down because we see only moments when the light is bright.
Same principle we can apply to roulette ball.
Instead of light we can use a timer that produces vibration every let’s say 1.2 seconds.

For this experiment let’s assume the rotor is not moving.
If we look on the picture point “A” let’s say it is the point where we see the ball at our first signal (strobe light or vibration). The ball is rotating it makes 37 +14 pockets in our strobing time of 1.2 s.


On next signal we see the ball above number 10, for us it looks as the ball only made 14 pockets.
Of course the ball is slowing down therefore until we get signal again it will not make 37 +14 pockets but let’s say 37+6 (8 pockets less). If we add from point “B” 37+6 pockets we will come to point “C” above number 20 and it is 6 pockets distanced from point “B”. When we get next signal the ball is at point “D”, the ball made (37+6)-7=36 pockets and it is above number 1. This time the ball is the closest to our previous point. Only one pocket difference therefore this is our triggering point.

In this example you can notice that the ball is slowing down 6-8 pockets in our reference time of 1.2 sec. This is something most realistic to happen on a real wheel in casino. I would like you to notice that with this approach we do not get a perfect crossing pattern match. As in this example in last strobed rotation we were one pocket shorter from point “C”. We could have plus or minus 4 pocket difference. When we find the closest match to our reference time where in that time the ball makes closest to one full rotation we know that ball time is close to our reference time, therefore we assume remaining time that ball will travel is constant.

From the moment when we found that ball rotation matches 1.2s time we wait until ball comes to our dominant diamond DD and read reference number from there. We assume that after that point the ball will make same amount of rotations until it drops therefore if look from the end of spin our reference number should be taken in same rotation. It will not happen all the time, the reason is + or - 4 pockets difference in our sample, we can avoid some spin or we can live with that and still play, but there is additional problem.

Firstly let me explain something! Here I am trying to explain reality of visual perdition when trying to target particular ball rotation. Same as with all posts in this thread, I do not have intention to discredit anyone who sells any of this or similar methods. It is only constructive criticism and personal analyses of problems we as advantage players are facing. Perhaps through all this thread I may have intention to explain why myrulet.com visual prediction 2 (VB2) is as it is, but we will come to that point later.

Back to additional problem with finding pattern using timer of 1.2 sec.
On the same picture notice the position of dominant diamond (DD).
In our example we found the closest pattern at number 20 -1, notice the position in relation to DD.
After we found the pattern, the ball needs to travel to our DD where we read reference number. It takes time, in this example it is probably about 0.5s. Therefore even if we found perfect pattern, real ball speed at DD will not be any more 1.2s but slower by how much the ball slows down until it makes about one third of rotation (on the picture it is from number 1 to number 0). If we analyse it more we can see that we could find same pattern at any position and not only 1/3 distanced from DD as in this example. We could find it after ball just passed DD; we could find it right in front of DD.

It means sometimes we have to wait almost for a full ball rotation until the ball riches our DD. Surely then at DD where we take our reference number, the ball will not be 1200ms but closer to 1400ms per rotation . Alternatively if we find pattern just after the DD, we may take reference number instantly estimating which number was at DD when the ball was passing it.

Basically if we find pattern in first half of the wheel from DD point, we can go back to DD if we find it in second half we wait until the ball finishes rotation as previously explained. This way we target ball speeds somewhere 1100ms -1300 ms, if we waiting for ball to come to DD from any position when we find pattern then we targeting ball speeds about 1200-1400ms. It really doesn’t make much difference if we do not know at which speed is a braking point when the ball will make one less revolution. Another option is as soon as we find pattern when the ball is 1.2 sec. regardless where we find it we instantly transfer eye sight to DD and read the reference number. It would us about same results.

Does it sound complicated? It isn’t complicated I would say more inaccurate and to improve accuracy it may be complicated. If we take in consideration that finding cross pattern perfect match is not perfect every time also that when the pattern is found ball position may be at different place we can assume differences in prediction can be significant. We can see problem of using continuous pulsing timer (thumper) to find cross pattern.
Instead of observing ball in reference to roulettes frame we can look to match pattern with moving numbers on rotor of the wheel as Laurance’s CS) , but differences in rotor speed will give us only more errors and make finding patterns less accurate.

For this example let’s say during single spin few ball rotations have times as follow;
1000ms, 1200ms 1400ms
With our timer we are targeting to identify when the ball is 1200ms
We learned that we can find pattern about plus or minus 4 pockets on each side.

Sure if we find it 4 pockets away, we do not have to play that spin, we can accept to play only spins where it’s 2 pockets away. We will continue analysing it with up to 4 pockets differences in finding the pattern.

If we find pattern and in our reference time the ball made about 4 pockets more (37+4) It means that the ball speed is faster than 1200ms by about ½ of rotation which in this case is about 1100ms per rotation.

If we find pattern and in our reference time the ball made about 4 pockets less (37-4)It means that the ball speed is slower than 1200ms by about ½ of rotation which in this case is about 1300ms per rotation.

Now we know that when we find pattern the ball speed can be in between 1100 and 1300ms par rotation.
But each of patterns can be found at any place, so sometimes we may wait ¼ of rotation until the ball comes to DD sometimes it can be ¾ of rotation. It adds extra width of rotation.

If we ignore and do not notice anything, we can have additional 200ms deviation added to previous one. So we are 1100ms up to 1500ms. It doesn’t sound very accurate. Since in this example difference in between rotations is 200ms we can see that our accuracy of predicting in same ball rotation is 2 rotations wide. Having it within 2 rotations doesn’t sound bad , but look the length of rotations . Two rotations together in time may take 2.5 seconds. Having so much differences in time until the ball drops will create a lot of errors. If rotor is 5 sec per rotation (very slow 7.4 pockets per second). We will float with our prediction 2.5x7.7=18.5 pockets. With faster rotor error will be higher.

This shows us how important for someone who plays such method even he uses timer is to observe where the pattern was formed, and how close to perfect matching pattern was. We can’t forget that finding cross patterns many times can be in area of the wheel where player can’t see it. Perhaps with some practice the player can learn to adjust.

But there is more about timers.

Now when we know all problems in finding perfect cross pattern, let’s look if it can be done on a better way.


This time instead of strobing the ball we will use only one time interval of 1.2s.
We are guessing when during the spin to start predicting, we do not need to be so accurate.
But we want to start timing the ball when it is faster than 1.2 sec, but not so much faster.
If we get signal and the ball is in position “A” , we wait for timer to give us another signal after the time of 1.2s is elapsed. It can happen at any place and we are looking how far that new position is from our first point “A”.

It can happen that the ball made 37+5 pockets or 37+14 (point" B") or 37 +20 (point “C” and “D”).
I wouldn’t go for more than 20 pockets extra, it means we started to early during the spin.
I hope you understood that we can have and anything in between this samples.
We can have ball in 1.2 s makes 37 + any of (…0,1,2,3, …18,19,20,21…)
Of course if the ball made 37 +1 we are in our targeted rotation of 1.2 s.

If we notice 37 + 10 we know we need to wait one extra rotation then take sample.
We can look at it as (37 + X) where X is actually indicator of ball strength.
In our example we have had differences of 6, 7 or 8 pockets in between rotations.
Knowing it we can apply it here. (But it doesn’t have to be on every roulette wheel).

I know many things from FFA roulette computer which can tell me differences in between rotations at particular moments during the spin, but there are and other way to know it. Something is explained in forums VB2 section but for example the player can simply map results and compare it.

Some who have FF/FFZ roulette computer already have adjustable timer from 200ms to 16 s in steps of 200ms.
This timer at any moment can be triggered by switch.

We can apply previous process, it is easier then to spot 2 times where the ball is, as it was the case of free running timer we described earlier. Especially considering that in some situations we will not be able to see the ball. When we have switch control timer we can start it when the ball is at DD and follow from there. We can adjust timer to any time, we used 1.2 sec only as an example.

If we use this way we need to exit prediction before then the knee point start declining and where such kind of prediction would start losing linearity.
Because we do not look for cross pattern as in previous example, now we can observe it on a different way. We can look how much extra the ball makes across rotor numbers in our time of 1s.

For example we start timer at 12 o clock and we have there number zero. After 1.2 s the ball is again at 12 o clock. If we look at it as in previous example we have 37+0 .

This time we going to look how much the ball have made from starting number zero.
Since in same time of 1s rotor moves as well, if it is 10 pockets per second it would be 10 pockets movement.
Based on previous example where the ball slowing down about 7 pockets per 1.2s, it is 5.8 pockets per second.

5.8 pockets is 15% of full 37 numbers and one rotation. 15 % of 10p/s that rotor travels, is 1.5pocket. We can add 1.5 to 5.8 and we get 7.3 pockets.
7.3 pockets is how much ball slows down per second compared to rotor movement.
For example we start timer of one second and after 1s the ball crossed 47.3 pockets.
Instantly we start timer again and the ball will make 40 pockets. Did you get it?
We started timing the ball 1 sec later and the ball made 7.3 pockets less.

Nobody ever said we need to use timer 1.0 or 1,2 s as a reference.
So why not to use time that most matters to us.

Since rotor moves 10 pockets per sec, if it is the most, reasonably steady speed we can
divide 10p/s by 7.3 pockets and we get 1.36s .
Without any problem we can adjust the timer and use 1.4s.
Now if we apply that time at “any” moment during the spin on left side of knee point we will always read same number as we are predicting in a particular ball rotation every time.

The reason for that is decreasing amount of pockets ball makes how it progressing during the spin matches the amount of pocket that rotor makes in same time.

Bingo, we come to myrulet’s visual prediction VB2.

With VB2 we can predict with same accuracy in let’s say 6,7,8,9,10 rotations before then ball drops,
Vb2 is not sensitive as cross patterns,
How does it compare if we do not use timer?

Well if instead of 1.4s we by mistake we used 1.3s we will read reference number 100ms earlier, if the ball is crossing rotor 50 p/s we would be wrong by 5 pockets.
If we have 100ms error with cross pattern or if estimating ball speed, we are wrong by full rotation.
With VB2 with 2-3 spins the player without any calculation can know which reference time to use but I will not be explaining it here since it is already explained.
The point of all is if we use time why not to use time that matters.
We still can observe how much ball makes in our reference time and compare results, but with VB2 we project it directly to rotor and we can read predicted number instantly.

Yes there is more

When wheel has a dominant ball drop point we call it tilted wheel even dominant point may not be created purely by tilt.

Tilted wheel groups different ball speeds to drop at one point.

If ball rotations of one spin are 1000ms , 1200ms and 1400ms and the ball hits our DD, it may be
1000ms ball makes 7 rotation until it drops , 1200ms ball makes 6 and 1400ms ball makes 5.


If we targeting to have prediction when the ball is 1200ms per rotation we know we would have remaining 6 rotations until the ball hits dominant diamond.
Previously we said, if we always can have prediction in particular ball rotation from the end of spin we would have remaining ball traveling time constant.
As much as that is correct it is also incorrect.

Few years ago I start arguing with some VB players, (especially case when rotor speed is fast), trying to explain my point of view.
The dominant diamond groups more ball speeds.
When we observe ball, trying to predict it when it takes 1200ms/r, the ball time per rotation will not always be 1200ms. It may be 1201, 1205, 1250, 1199, 1190, 1150ms or anything in that range and the ball still can make 6 ball rotations until it hits the DD.

But my argument is will the ball traveling time be still the same for each of ball speeds?
No it will not. One ball from 6th rotation may travel faster until it hits DD and the other one will travel same distance in longer time. We assume that on a roulette wheel faster ball will travel longer. But on a tilted wheel it is actually opposite, faster ball within one rotation frame will travel less.

This graph uses data from real roulette spin, but some data is added to display what is happening if ball is a bit faster or slower from data clocked inside computer.
If we look point A red dot, we have ball of 0.93s per rotation speed that travels exactly 11 sec until it drops. Green dot is still within same rotation until the ball drops and it tells us 1.12s ball will travel 12.1s until it drops. (This is based on an assumption that the dominant diamond gets 75% of hits.)
As you can see, slower ball to complete same amount of rotations travelled 1.1s longer.
Now if we look point B on the graph we can see that time deviation is reducing when predictions are closer to the end. So a bit later during the spin prediction can deviate 1s or, again later it can be 0.7s…etc.


This tells us that if an advantage player every time predicts correct ball revolution he is still wrong by some amount of pockets, since the ball speed is not always same therefore remaining time to the end of spin changes.

If ball’s remaining time floats let’s say by 0.7s (usual case) if rotor speed is 15 pockets per second prediction results will deviate by 0.7 x 15 =10.5 pockets. It also tells us a VB player is better off if he play wheel where rotor speed is slower. If for example the wheel was 9p/s then possible errors would be 6.3 pockets.

Yellow mark on the graph is also interesting. If we look from green point, the ball time per rotation did not change much but remaining time for how long the ball will spin changed almost by 2 sec. (For someone who has FF roulette computer this explains position and importance of “point set” PS)

Sometimes the error is not error if it gets compensated by another error.

If rotor is slow, 6.3 pockets error when the ball arrives to DD I’ll be compensated.
Ball that travels longer will also hit DD at lower position therefore it would drop on rotor more vertically and faster. Stronger ball that comes in shorter time to our DD will hit top part therefore it will fly (8-9 pockets) from DD until it hits rotor.

Deviation of ball traveling time explained in previous post cannot exceed length of rotation where we have prediction, other way it will be an extra rotation.
From all of this for the most accurate prediction we come to conclusion that most what counts is ball remaining time regardless of how many rotations the ball makes to the end.
Yellow mark on previous post graph is very interesting. If we look from green point, the ball time per rotation did not change much but remaining time for how long the ball will spin changed almost by 2 sec. (For someone who has FF roulette computer this explains position and importance of “point set” PS)
It is logical at green point the ball has the slowest speed which is distanced x amount of rotations to the end.

At yellow point the ball makes one rotation less (x-1), but also the ball makes remaining rotations in the shortest period of time. At this point during the spin only 70ms difference in ball time per rotation creates about 2 seconds difference in ball traveling time. To avoid this, a classic VB player for this example will need to have skill to predict when the ball is within 0.93s and 1.12s. First of all, the player will not have clue about this figures and where are limits of one ball rotation, next it is very questionable can he define ball speed accurately. Without timer, counting or using rotor that can also deviate in ball speed it is a fairy-tale. Even with timer it would be close to impassable. All what player can do is to accept possibilities of errors and play. His predictions may look as this (rotor adjustment is excluded).


When he predicts within right ball rotation when the ball speed is about one sec. per rotation his accuracy is represented with green area. If he misses ball rotation his prediction will be wrong and deviate within blue area. If prediction is later in time, collared areas will be narrowed but difference in between green and blue area will be greater. With decreased rotor speed, collared areas are narrower and closer to each other. If there is rotor speed change and the player can make proper adjustment, even as it is the player has a reasonable chance to play with an advantage and win.

Most of the time hits on blue and green area will not be equally distributed, for example 70 % hits in one area even it if is in blue area (error). will give to the player more advantage since he will make an offset adjustment and consider blue area as his main target. Do not forget that this is only when the ball hits DD and if his accuracy is within 2 ball rotations ~2sec of time interval.

VB2 might be worst then that, but I like it because it has some other advantages.

VB2 predicts earlier in time, we have more time to place bets, we do not have to look so much in to the wheel ,therefore we do not take much attention as an advantage player.

With VB2 is requires us to target ball speed before the knee point, so we can say 0.8s/r iis good time.
When ball rotation is 0.8s, the rotor that travels 10p/s makes 8 pockets.

With VB2 we predict in any ball rotation therefore we are never wrong to miss particular ball rotation. But we may always be wrong by some amount of pockets since prediction gradually increases proportionally to ball speed. For example ball of 750ms/r can be predicted as zero, ball of 760ms/r may be predicted as 26, 770ms/r as number 3 …etc (ball in clock way direction)

Slower ball gives us shorter prediction. That way system follows rotor changes, but previously we said that at some points during roulette spin slower ball travels for longer. That makes additional errors for VB2 system. Prediction changed from zero to 26 but in some occasions it could be 32. We do not have prediction in wrong rotation. VB2 instead of constant amount of remaining ball rotations is after remaining time regardless how many rotations is there.

This picture approximates errors with VB2. It is similar as if it is predicted bit later with classic VB on the previous graph but here we have just one area. Each area gets reduced with diamond deflection but with traditional VB we still have a gap in between 2 areas. Traditional VB player can be more accurate as long if he can keep prediction always in same ball rotation. Some people say they can.


But VB2 has additional advantage that relates to real casino environment. It really doesn’t need much from the player, it is more relaxed way of play, but the main advantage is in total results.

Roulette wheels are never 100% tilted, the ball doesn’t hit DD all the time.
What happens when the ball hits another diamond?
Our advantage is reduced to negative so we have more chance to lose.
That is the truth, but it is not always the case.

Imagine DD above zero, clock way direction ball (CW) hits there and drops at zero.
But what if the ball hits diamond positioned at 3 o clock?

If the ball makes extra ¼ of rotation it will make spin about 0.5s longer. (Last rotation about 2 sec. long), the ball will hit at 3 o’clock but in same time rotor will move so the zero will be around 10:30 o’clock. Result is completely wrong.

If the ball makes extra ¾ of rotation the spin will be ~1.5 sec longer.
So the ball will hit at 9 o’clock diamond, zero will move by about 15 pockets in opposite direction so it will be at 7:30. As you can see it is close to 9 o’clock, we have 5 pockets distance error which still may produce us an advantage.

To have this perfectly balanced we need rotor speed of about 6 pockets per second that is a rotor speed that takes about 6.2 sec per rotation. It is very slow rotor speed and we do not find it so often in casinos.

VB2 with increased ball speed predicts more pockets.

Until ball speed reaches moment where it may go for additional ¾ rotation, VB2 already added extra pockets to prediction, therefore instead of zero it may predict 21.
Number 21 would be more close to 9 o’clock dimond hit (as in previous example).

Same explanation applies if the ball makes ¼ less of rotation (no advantage) or ¾ less of rotation (still advantage).

The point is that if we play wheel where ball is hitting other diamonds as well we get some advantage even when the ball hits different diamond, with VB2 rotor speed can be faster than 6 pockets per second. More realistic rotor speeds is ~4 seconds per rotation, as it is at places where i play :slight_smile:

Good post but i have one main problem. I am a casino dealer and we need to know the last few revolutions in order to call “no more bets” and do our job . we don’t need to use fancy techniques to figure out the final 4 revolutions.

yes but in ur job u don t care if the last revs in the time that u say NMB is the 5th rev or the 4th or the 3d…
with VB, 1 rev mistake is a desaster!

But I agree that in some wheels (that have a strong knee point) we can find the same rev in a very good persent of the times.

For dealer it is not much important if NMB is 3, 4, or 5 ball rotations to the end but for a player it is.

Sure anybody can reasonably well estimate, in this thread that’s explained as well, also it’s explained at which point during the spin someone can estimate it the best.

If we put it all together, someone who tries to understand it, most likely will come to conclusion that roulette visual prediction is something very individual. Most important to understand is principles of tilted wheel and that’s explained for FREE at many places on the net.

If we look systems to buy, it look as for more money we pay; we get less.

I’ll just touch few examples.

We have heavily advertised “genuinewinner” system sold for $2000-$3000

When someone buys it, all he gets is close to basics of dealer’s signature and empty promises.
After many people who purchased it defined it as a scam the seller start changing it, claiming that now he includes and Laurance’s system as well therefore it isn’t a scam. Also the GW system has option where players can sent him 333,666 or 999 results from one wheel and apparently by some magic he discovered in nature he will be able to tell them which numbers to play. (Wrongly interpreted bias wheel play)

Additional system on market was Bob Gordon’s system for $1000, which is a bit better then Genuine winner system but still it’s is still far from promised.

Then there is Laurance’s systems

Professional Roulette Prediction: Volume 1 - Basic Methods $250 or
Professional Roulette Prediction: Advanced Methods, with software $500

Which doesn’t come with any promises but it is quality explained document with original ideas which obviously first 2 explained systems do not have. (Apparently even some of FF roulette computer users use Laurance’s software.)

Jafco’s Vibe is about $300, it includes explained principles of tilted wheel and a timer that generates vibration every 1 sec.

I believe esspri trying to say ,why systems, why timers, why all that theory, if someone as him simply can estimate reasonably well. And he is right but he still needs to know what to estimate and why everything is as it is and how to increase chances. But yes, someone who understands tilted wheel principles doesn’t need any system since he can make one for himself.

No systems high price can be justified, perhaps Laurance’s and Jafco’s systems are close to justified, because people get some original ideas, software or timer.

Surely, any copycat can repack VB principles and with big advertisements on internet find naive people to sell it.

The best product on the market is FF. :-*

Wait, we did not come to that point 8)

I prepared here five examples about different possibilities at roulette table. With each example we play with an advantage.

Example 1

We play 66 spins, number 66 doesn’t have any particular meaning it could be any number, because I started explanation with advantage of 1 in 33 hit rate, so I used 66 spins so an average hit on a single number would be 66/33=2.

Also around that figure it may be amount of spins that a player may play in casino within 2-3 hours’ time frame.
In all examples we will play $9 per spin.

Example 1 and 4, we play $9 on a single number while in other examples we distribute $9 across 9 numbers sector, so we place $1 on each number.
If we play 66 spins total amount placed on table would be 66 x $9 =$594
with an average hit rate 1:33 we would expect to win from table 2 times

2 x 36 x $9 value = $648
Place $594 get back $648, profit $648-594=$54
As you can see it is close to 10% profit which is logical since we have an advantage of about 10%.
However, because of short sample and unequal distribution of results we can expect something else.

If we look possibilities what will happen by playing single number we would 95% of time end up in between minus 0.79 and 4.79 wins. Of course it is a mathematical limit, we can’t win 4.79 times, it would be 4 or 5 times and we can’t have limit lower than zero therefore minus 0.79 is zero wins.
After calculation using possibility limits 95% of times we will end up somewhere in between winning 0 to $1550 from the table we play.

It means we will end up losing every single spin or we may profit up to $1550-$594=$956.
Surely such high deviation is not very good for an advantage roulette player.
As you can see with 10% advantage we can play 66 spins without single win. The ball may stop 3 times next to number we played but it doesn’t make anything good to us.

Example 2

Here we have same situation but this time we distribute same amount of money $9 across 9 numbers and place $1 on each.
We expect to win 2 times per number played which makes an average of 18 wins.

And again, due to uneven distribution 95% of time we will end up from 10.76 to 25.24 wins in our 66 spins played.
With 10.76 hits we will still lose since we would get from table $387, and we placed $594.
As you can see this makes distribution more even, after 66 spins 10% advantage 95% of time we will end up in our hands with $387 to $909.

Example 3

To be 95% sure that we will profit when playing with 10% advantage we would need to play same as in example 2 but about 1500 spins in total.
To have constant advantage of 10% for 1500 spins using ballistic is extremely hard.

Some may say well dealers signature on long run may show us advantage of 5-10%.
The dealer may spin 50 spins 15,16 or 17 rotations most of the time.
Next 50 it may be 16,17,18. Next 5 mostly 17, next 50 mostly 15…etc
Rotor may deviate in speed, 1-2 pockets per sec, which on spin 15 sec long makes 15 -30 pockets differences. Ball may behave differently, and it does. 50 spins ball may be jumping differently.

The player without constant observation of ball drop point has no control and no idea if he is still playing with an advantage. Results combined with randomness of ball jumps are more pattern chasing, which we never know when will start or end, or how long it will last.
With small advantage for an AP everything is harder.


[i]Anything above green line is advantage. If we read values we can see that pick point is only 2.85 and green line is at 2.78 so the difference is 0.07. It means that for every 100 spins played we would win 35x0.07=2.45 units. Or if we round it up, we can say every 200 spins we should profit 3 units. Theoretically it will work; you can play for 10 hours and win few bucks.

The problem is that hit rate is only 0.07 above the profitable limit. Therefore we wouldn’t be able to detect it. It will take hours to define it but until then definitely the dealer will change. He may start spinning more often slightly different amount of rotations, the ball will be replaced or to be more dirty, polish on wheel may erode, temperature or air pressure will change, the wheel will slightly move from vibrations on the table, the tilt will change etc. It means soon we may play with negative advantage.[/i]

Example 4

It is same as example 1 but this time we have about 30% advantage (hit rate 1:24).
If we play only a single number we still can end up without single win.
We will place on table $594 and 95% of times we will end up with getting from table 0 to $1943.

Example 5

It is same as example 4 but this time we distribute $9 across 9 pockets ($1 on each)
Since advantage is significant ~30% and units are distributed across 9 pockets 95% of time we will win. We would place on table $594 and we would be getting back something in between $608 and $1174.

It is important here to understand that just playing more numbers doesn’t create same effect all the time.
Someone may come in conclusion that playing 18 numbers will give us better stability and certain wins. It is not the truth since created advantage is not across 37 numbers. With each spin one area of wheel will have less chance for ball to stop there, while the other area will have more. It depends how wide area with advantage is. In usual I play 5-9 pockets. If playing wider sector some numbers we play may have significantly reduced advantage or even negative.


This may be typical created advantage.

Everything marked with green colour is where the advantage is. After analysing data if someone was playing according to this prediction and covering 11 pockets from -5 to +5 from prediction, he would play with advantage of 53%. Since this graph are results from bench test, in casino we often can’t analyse results with such details. Therefore playing 5-9 pockets will give us some space for errors if we are not perfectly positioned in area where the best advantage is.
It is important to understand how advantage for ballistic AP player is created.


To be able to have previous graph the player will need accuracy of prediction to rotor point as it is displayed on this graph. We can see that most of spins the ball drops within 10 pockets of players expectation. After ball hits the rotor it jumps based on ball scatter law, for that particular wheel and ball. When a player has something so constant to observe, he has better control over the game.

From the graph we can see that the ball most of the time hits 10 pockets in front of predicted number.
If for example for next 5-6spins the ball hits 7-8 pockets after the predicted number the player instantly knows something is happening on the wheel.
If the player observes only results where the ball stops he would need at least 100 spins to be able to react and to make necessary adjustments so he can continue playing with an advantage. Playing 100 spins and losing is not fun.

That is why I was shocked when come across roulette computers that build data graphs on that way, or when watching some salesman making videos where computer can’t constantly predict where on rotor the ball will drop.

Anyway if you read this post with understanding you would also understand that making video of 50 to 100 spins in convincing someone to buy roulette computer or VB system is pointless. :smiley:

95% of times this is our chance to end up with any profit with each of previous examples.

Example1 … 53%
Example2 … 60%
Example3 … 99%
Example4 … 64%
Example5 … 100%

Forester i have to admit i like your effort very much and this post is great, thank you.


[quote=“lucky_strike, post:19, topic:538”]Forester i have to admit i like your effort very much and this post is great, thank you.


Regarding Foresters assumptions that when ball divide from fast to slower is not accurate - what to say then take a hike and educate your self :slight_smile:

There is no difference using head count comparing to visual errors and feel with thumper except with thumper you know for sure what interval you have from read to smack and develop a head count with 0.9 1.0 1.1 is based upon personal loop and visual observation …

Example easy head count with visual ballistic observation.


From number read below deflector to ball strike deflector with 4 to 5 pockets error.

Time 9.5 9.3 9.1 9.0 9.2 9.2 !!! compare that with VB2 and this comes with method to adjust with one pocket difference reading rotor and adjust fudge factor …