# Understanding roulette wheel prediction

Recently I have received an email from friend of mine. In his email he referred to Stefano Hourmousis new approach of discrediting the FF roulette computer.
It isn’t really much new, it is only who knows how many times edited version of same page created a year ago.

I better use it and explain what really the roulette prediction is.

For start I will quote Stefano’s words where he is referring to" Forester’s roulette computer test" which is basic indicator will computer work or not.

Stefano says:

“His test sounds good in theory, and it would be a good test if a wheel was “perfectly” level, but such wheels simply don’t exist.”

Stefano is saying that my test isn’t real. For him taking a video spin from real wheel, clocking it and getting results which can be compared isn’t real. Then repeating it and still getting same result, then clocking the ball at the other position (it is still the same wheel which may be slightly tilted) and guess what, computer took different speed in calculation combined with different rotor position, it took all clocking errors and still produced accurate result that belongs to that spin.
For some reason this test isn’t good for Stefano. Further more same test applied on his wheel that he calls semi tilted produces same accuracy. In his video he claimed that he can’t see the ball at the moment of zap but all 21 spins were predicted with high precision regardless his wheel have had slight tilt.

Instead of my test where the ball and rotor are really spinning he suggests that in his mobile phone you type in for example 1000ms and you will get result. Woooo that’s interesting. Only he doesn’t point to what you going to compare that result. And he doesn’t tell you; when you type in 1000ms it doesn’t have to be 1000ms when you trying to clock the ball of 1000ms per rotation.

It’s hard to understand how someone can claim that typing in something where you don’t have anything to compare, or where you even do not know will the “phone” measure time accurately, can be as good as real time test where you have result to compare with.

But Stefano has to do it that way. The reason for that is that mobile phone most likely will measure time in ~25ms intervals, and even that may be worst since mobile phone is running additional applications in background. Which of course will have higher priority then his measurement?
With attempt to transfer my computers software to mobile phone, all five professional developers told me, it is imposable to get such accurate timing with mobile phones or PDA’s.

What really happens, if time is measured under OS as it is case with mobile phones or PDA’s?

Phone may display time in ms but since refreshment of processor is in about 25 ms intervals we simply can’t get accurate measurements. It means that measured result of time can deviate by 0 to 50ms and it will average to about 25ms.
That definitely isn’t god enough for roulette levelled wheel prediction and hardly it may be good for tilted wheel. The FF measures time in 0.001 ms and it doesn’t run any other applications, all hardware is working only to get accurate prediction as much as possible.

What surprises me more is why Stefano says that slight tilt will make prediction worst only for the FF. When at his video which comes together with his mobile phone roulette computer he clearly states that even when roulette prediction is done on semi tilted wheel his computer assumes that the wheel is perfectly levelled.
Stefano only follows what I write at my forum as potential problems with roulette prediction and then claims his computer does everything (as in usual).

When we are clocking the ball we are doing it at one spot and we can’t have idea how the ball behaved in between clicks ¼ (or ½ or ¾€¦etc) rotations from there.
Even in such case the FF is better, since clocking can be done at any place and completely independent from rotor. Therefore the system can learn from any position on the wheel.
But in general we are taking timings of full rotations we do not know where is slight tilt therefore the computer definitely can’t do anything in regards to that. If we know where the tilt is then we can use tilt system prediction.

Stefano’s scattergram can’t help, since things on the wheel change and his results are anyway combination of ball jumping across the rotor. We all know that the ball jumping with high percentage is random so it is pointless to use it to define something that requires few ms precision.
If at one particular spot the ball speed is 1000 ms and 9 pockets from there if it is 1050ms, tilt may cause it to be 1048ms. Definitely we can’t use 2 ms difference to shift final prediction by amount of pockets that tilt would cause in last rotation when the ball is due to drop. Especially not if time measurement is done with mobile phone or PDA as Mark, Stefano or Barnett are using. It is same as taking schools ruler to measure thickness of human hair in microns.

At his page Stefano grouped wheels in this order;

I. Heavily Tilted (Less than 4% of wheels):
ii. Semi-tilted (About 95% of wheels):
iii. Perfectly level (less than 1% of wheels):

I would say perfect wheel doesn’t exist; same as he did not write perfectly tilted whee but heavily tilted. Maybe we should use term reasonably levelled roulette wheel.

Expert visual players can take advantage if wheel is only slightly tilted. For example if one area instead of 25% of hits gets 35-40%. Therefore the VB player would really play semi tilted wheel. Such wheel is also possible predict successfully with IQE6. If hits are 50% plus I would probably use tilt system because it will produce greater advantage.

However, properly adjusted tilt system can be applied and on wheel with only 40% tilt.

It isn’t rocket science that if we have more tilt that we may get better result.
So really we can apply levelled wheel prediction to at least 70% of wheels where remaining would be heavy tilted or with significant tilt where we may be better of applying tilt system.

It isn’t secret any more that Stefano is selling levelled wheel prediction computer where he is actually trying to get advantage of slight tilt expecting that it will give him some advantage. Same as he believes in his genuine winner system. His computer doesn’t have needed accuracy even for that nor will his GW system provide enough so the advantage can be recognised and used to win.

Stefano’s computer can’t even handle rotor speed change, for any significant change his computer says risk and doesn’t predict. How then he can handle the ball measurement and calculation?

The FF has high precision so it can learn from spins while with Stefano’s computer you first need to visit casino with small camera that he sells for \$800 (only 30 frames per sec) then go home use video editing software to read timings in 40 ms intervals (yes that is how video is recorded 40ms is accuracy, but we can slightly approximate values in between) enter details about that in computer then go and play. And he calls it “high technology”, “hybrid computer”.

Did you ever see him writing all of this? Why not? Instead of that he is writing about me and my computer problems. Stefano dedicates complete web pages talking about me and there is no need for that since if you read my forum you will find out that I write much more about problems then anybody else. All people who own my system are welcome to do same and most of them helping me to make it even better and doing their own experiments. More confidential subject are kept inside support forum but there is and plenty for visitors.

Further more at his page Stefano writes:
A strong tilt makes a wheel very easy to beat, and a perfectly level wheel would still be relatively easy to beat with the right technology. But the problem is just about every wheel is somewhere between level and tilted. That is in order for most roulette computers to produce predictions that are accurate enough to overcome the house edge (and be profitable), most wheels are not tilted enough to be treated as a “tilted” wheel, and not level enough to be treated as a “level” wheel. Most wheels are “semi-tilted”.

With this part I may agree, but he is saying something that he really doesn’t want to say.
The FF roulette computer has right hardware to produce advantage and mobile phone definitely doesn’t. The FF measures time with tolerance of 0.000001 seconds and mobile phone in 0.025 s.
So about which technology and accuracy he is talking about? The only way to use mobile phone is to remove all software remove firmware and to program it as the FF is done to be able to do accurate time measurements.

Next from Stefano;
“Another problem is many wheels have ball track deformations, which are likely to affect results.”

I think he should bring this up when he was selling his phone instead of promising people to win at any roulette wheel. It must be that he didn’t know about such effect since he was advertising “win at any wheel”.
On some wheels deformations may be so intense that prediction is imposable. It really doesn’t matter how accurate we measure speed if once ball of 1000ms per rotation travels 5 rotations and next time the ball with same speed travels 6.5 rotations or anything inbetween. Such scenarios are real and I was talking about it from the start. But that doesn’t affect only my roulette computer but every computer is affected. At least with the FF time of prediction can be shifted so such effect in some situations can be reduced.
Once when we measure the ball speed and get prediction there is no way that anybody can know how that ball will be affected by deformations on ball track, by every particle of dust, by table vibrations etc. We can predict particular wheel only if for example measured ball of 1000ms most of the time ends up at same place with reasonable tolerance.
The higher percentage of grouped results we have, the highest advantage we will have. But same spins doesn’t have to take ball to same spot every time (according to measured ball speed) and there is nothing that anybody can do about it.

Anyway, even in that field the FF is better designed. With Stefano’s computer the player estimates when he will have prediction so it can be once at 5 sec to the end next time at 10 sec.

What if the wheel has deformations and if only possible accurate prediction is when the ball is under 8 sec to the end.

With the FF player can specify it and the system will predict at that time. What Stefano still doesn’t know is that there is a sudden change in ball deceleration at some point during the spin. Some people who analysed it (much earlier then me) call it “the knee point”. Prediction is always more accurate if it is done right after that point. FF can do it but user of Stefano’s computer can only guess where it is and estimate time when to clock the ball.

Recently I come across part of “Mathematic of Gambling” By E.Thorp.

I found his work interesting. His research is similar to mine and he confirms that ball with particular speed will most likely end up at same place. However Thorp never finished the project. His missing part was practical application. He never solved problem how to measure accurately ball time to be applicable in real play.

That was actually the first thing that I have done. Without that I wouldn’t work on the project.
Whoever used the FF would know that the system instead of pronouncing number gives signal to the player when the ball is over predicted number. Now imagine which kind of calculation and which precision is required just for that. Other words it means if the FF can calculate time of two different speeds decelerating objects, when they will meet to display result, it must have accurate error correction and calculation to do it, or when the test is applied results would be all over the place. If such precision and accuracy isn’t enough to predict roulette then nothing else would do it.

Some people believe in magic voodoo or some similar orthodox approaches.
A year ago someone gave me access to see what Stefano’s genuine system users need to deliver to Stefano so he can respond back to them with how to play particular wheel.
Stefano claimed that he discovered some kind of magic and natural numbers in nature. So he asked them to collect only particular numbers of spins in amounts of 333, 666 or 999. I do not know if he still have it but it was something that really shocked me. I can’t comment magic but I definitely can say his computer by physics, Newton’s law, software solution is rubbish. And it was rubbish to all people I know who bought it. If he wants to claim that his computer can find out by nature numbers or some orthodox parameters why once ball of 1 sec will travel 6 rotations and another time 6.5 I can’t argue with it since I do not understand dark forces of magic.

Stefano was imprisoned. I wouldn’t look on that so bad if it was for something else. But he spent time in jail for scamming people. Therefore and his voodoo prediction I classify as nothing else then scam.

For someone who already spent time on the net this is nothing new. But if you are new to this, then do your researches talk with people and you’ll find a lot.