Development of KIN computer; 1983. I am wearing the Black Jack computer while holding a spare. Also shown is the KIM-1 computer, attached to keyboard and monitor.
Photo Joyce Stonehocker.

 A      Letters

One of the teams I did find extremely genuine and their site worth looking at was here:

These guys are extremely interesting and genuine for me, the guys that have closely obtained similar results to myself, there is others on their site discussing similar projects.

These are the good guys with genuine research and products.

You will find their data and correspondence most interesting with fellow like minded people, including myself way back in 2002 with a post on their site.

I really admire this team, they had little in the way of technological advances in the field of mobile portable computing and so was limited to what they could achieve, but they achieved a substantial amount of excellent results similar to mine. These guys are at the top of the list for me in roulette prediction development for they exceeded where others would have failed with technology at that time.

 Link  Pictures of KIM-1

Sent: Tuesday, April 03, 2001
Subject: Question

Dear Sirs! Hi!
Predicting roulette is of great interest to me. I have looked through your data shown on the web-site.
Of course there are many questions about what you did but first determining question is :
It seems that errors according to timing of ball or rotor by means of switching timer are too big (too big
to have enough advantage) because of variable delays of human reaction (time between an event happens and
human hand switches timer ) .How did you take into account the errors according to unperfection of human
reaction during timing?
Thanx in advance.

We did not do anything about errors due to human reaction time. We realized this problem. I suppose some help could have been had by trying to adjust the errors by the program. We did get a fairly straight line, however, when timing a constantly rotating object; this made us think that the error introduced due to reaction time was not too great. We even worked on a device which could sense the rotation via photo sensitive diodes (built into eyeglasses, but we did not develop this.

Thank you for writing, Dimtry

Are you in Russia? I was in Russia last week! 
 Link Buran - Strar City - Space


Sent: Wednesday, April 04, 2001
Subject: RE: Question

Yes I am in Russia. It is an interesting fact that I have the same idea about using photodiodes.
If you don't object I would collaborate on the subject with you with great pleasure.

I also would be glad to meet with you when you are in Russia.

I only wonder why you have left such a perfect idea of
earning money by predicting roulette?

Photo Marina Matveeva ©

St. Petersburg, Russia, Oct 2000
St. Petersburg, Russia, Oct. 2000

Sent: Thursday, April 05, 2001
Subject: RE: Question
Importance: High

Hi, Dimtry
I haven't worked on this project for probably 15 years or so. At that time, I would have given an arm and a leg to have the small TV cameras which exists now and computers with megabytes of memory, for image processing. Where do you live? I juts returned from Moscow Last week and I was there in October, also Sankt Petersburg.

Sent: Friday, May 04, 2001
Subject: From Dimtry about roulette

It is Dimtry from Russia again.
Let me ask you some questions.
I need to understand exactly what means 'touch down
time' in your data.
Is it the time when the ball at first time hit the
rotor after leaving the upper track of the stator and
reflecting from "diamonds" on the stator (is it the
first touch on the rotor)?

Or is it the time when the ball entered the winning
pocket and stopped there waiting for a dealer to stop
the rotating rotor to claim the winning number?

Did you time the touchdown moment by reacting to
sound of the ball hitting the rotor or how?
--- BY sound and by sight.

Further, Is the target slot a slot (pocket) onto
which the ball hits when it falls from upper track and
have been reflected from diamonds (that is it is the
place where the ball at first touches the rotor)?
--- If no reflection target slot is the one it landed in. I f reflection, target slot is the one the ball "would have landed in", if no reflection had occurred --- by estimation.

Did you analyze the motion of the ball from the
time it left the upper track on the stator until it
crossed from the stator to the rotor?
--- NO. We analyzed "skip", the +/- distance (in units of pockets, or slots), that the ball landed in, from the target. I think there is a graph of that on the net.

Does this always take the same amount of time?
How may this time vary?
-- I think so, roughly, I don't remember.

Sorry for so many questions:)

Also, if it is not difficult for you, could you send
me more data for analysis.
--- you have all of our data already, sorry.

Thank you very much.

Sent: Wednesday, May
To: nowscape
Subject: From Dimtry about roulette

It is [**] again.
Excuse me for frequent disturbing you.
I have taken a look at Nathans patent you denoted at
your site. It is rather different project isn't it?
That is Nathans patent have nothing to do with your
project, is it true?
If so, what do you think about his view on the
By the way, I have found that in Russian casinos,
dealers change direction of ball after each spin (that
is one in clockwise direction, the next one in the
counter clockwise direction)? If American casinos have
the same feature then did you take it into account in
your project?
Thank you

Sent: Wednesday, May 16, 2001
Subject: RE: From Dimtry about roulette
Importance: High

Dimtry --
I doubt that the direction of makes much difference, especially on a wheel in which 50% of ball travel is in both directions. In US, I have only see the ball spun in the clockwise direction, but there is nothing that obligates the casino to always spin the ball ion the same direction. It would be interesting to see data comparing the two directions. My friend who worked on this project also says that Nathan's' patent will not work -- that there is some flaw in Nathan's logic. We have had many discussions over this, and I never believed that Nathan's scheme will not work. But my friend is very smart, and so I will have to make some allowance in my judgment. I often thought that Timing only ONE event would work as well as timing the ball and wheel (rotor) separately. That event is the instance in which the -0- or -00- on the rotor passes the ball, going in the opposite direction. This is similar to planetary motion, in which the more central planet overtakes a planet orbiting further out in space. If the near planet obstructs the view (as seen from the sun) of the farther planet. this, I believe would be called, in astronomical terms an occultation. That's what we called it, the Occultation method. I looked into his once and derived the formula... It is something like B*W / B+W; Correct me if I'm wrong. I've also considered another method, but we did not implement it. That was to do a simulation of rotating objects in the computer, with a control, that would speed up or slow down the rotating "objects", in synchrony with the real world. Perhaps audio feedback could be provided to signal that's happening with the relation of simulated rotation and with the outside. At some point, control would be relinquished to the computer, and it would carry out its simulation at lightning speed, in advance of the real world outcome. If you try this and if it works, you owe me 2 percent... that is if you haven't already thought of it or rejected it :)

BTW, I was in a casino in Sankt Petersburg in October. I paid little attention to roulette, so I did not notice the changing of direction, but it seems to me that this would have stood out like a sore thumb, to me. I did loose $60 on BJ -- they only deal 1/2 way down the shoe. I couldn't believe it!

Sent: Thursday, May 17, 2001
Subject: RE: From Dimtry about roulette

Thanx for quick answering!
I suspect your friend is right that Nathan's logic is wrong. The moment of falling of the ball from the
upper track down the rotor is determined only by the velocity of the ball relative to the stator (the
velocity at which the ball drops from the upper track is constant and is determined ideally by the friction
force between ball and track, diameter of track, slope of track) but not by the velocity relative to the
rotor (rotor has nothing to do with the drop-off moment)! If I am wrong, correct me.
As for the method proposed by you, unfortunately I have not understood the idea completely. I just wanted
to use 3d max studio for simulating the process of rotating of the ball and reflecting from "diamonds"
but have chosen to purchase a real wheel as less expensive (in time terms) variant.
If it is not difficult for you please explain in details your method?

As for St. Petersburg casinos I recommend "Alladin" penetration up to 90%, early surrender (but they hate
counters :)). In general, penetration 60-65% is common in Russia but I play only in games with ES vs Ace and
Thank you very much

Sent: Thursday, May 17, 2001
Subject: RE: From Dimtry about roulette
Importance: High

I think you are right, on that account. The time to drop from the rim is constant, and thus can be always be added into the calculation of the (predicted )time at which it will drop from orbit.
I don't know 3d max studio. When we did our work, we ignored the effect of diamonds, since their effect seemed to be more or less a random perturbation of the trajectory. See .

My friend's objection to Nathan's 1979 US Patent 4,357,659: betting calculator -- was not based on the problem you mention, as I recall. It was something else. I'll forward this correspondence to him, perhaps he will see fit to reply.

Do you have a casino-size, standard wheel? We purchased ours in Reno, Nevada circa 1994 for $3000. It turned out to be quite biased, favoring one segment of the circumference. It's the one pictured on the web pages.
What is ES vs. Ace and Tens? My Russian casino experience aw quite an experience, as they say. The building did not look like my image of a casino, from the outside. Inside, I had to go through a metal detector. The place was half casino and half disco, separated. At first the young man guard at the casino entrance, inside, would not let me in because he said I had "sport shoes". I had a girlfriend withy me. I feigned indignation and acted as if we were going to leave -- she did not want to -- and the boss came and he let us in. I guess casinos in Russia are for high-class people only, which I am obviously not.

90% penetration sounds quite good, this might almost make it worthwhile to visit St. Petersburg (and Olga again:). (I used to have (almost) free travel on Delta Airlines, but that has now stopped, I'd have to pay for the trip out of my own pocket.) I have forgotten where you live,
Dimtry, tell me again, if you write back, eh?

My simulation method was never developed in detail. It's a simulation of rotation, in real time, parallel to the (ball and wheel motion of) actual roulette game. at the time, during play. When an answer is desired, the simulation is allowed do run at the fastest speed the computer can run it, yielding a prediction ahead of the actual outcome.

Another method I tried -- using only one rotating object, a horizontally mounted bicycle wheel -- was to write a program which, without knowing physics, would predict when the wheel would come to rest, given an arbitrary initial push... This program ran on the same machine as the black jack computer. The idea was simply to map rotational speeds to predicted stopping time, and to extrapolate in case of "not enough previous data". One day I will have to post that on the net. My friend pointed out that the results I showed to him were no better than chance, and I suppose he may have been right. At the time it looked to me like I was on to something.

If physics fails in predicting roulette, I suspect that neural net programs might do better. IN the back of my mind, I've had a plan to try to fit the data I have into a neural net scenario. I have tried, unsuccessfully, to do this, with various free demo neural net program on the internet, about a year ago. Perhaps it is soon time to try this again, with easier to use software. If you know about neural networks, think about this, sometime and let me know what you thought :)

Best wishes,

Sent: Monday, July 23, 2001
Subject: From Dimtry about roulette


It is Dimtry again.
How are you?

For two month we have advanced a little!
We bought a good restored wheel. It seems not to have a preferred sector.
It is very hot in [Moscow] this summer. Every weekend I am tickering with the wheel collecting
data and process them on my computer.
I finally have received the book "Eudaemonic pie". I was pleased to see they had the same ideas I worked
out myself before reading the book! Although, it have rosen many questions. I think you read it, didn't you?
It claims, for example, that beside friction forces, wind resistance alone is mainly responsible for the
ball slowing down. On my calculations, the wind resistance or drug force invests only from 6%(at high
speeds) to 0.7%(before falling off) to overall slowing down. What do you think about it? Did you take it into
account in your algorithm? Also they considered the tilt very important. I don't know how to implement
it, although it may be important. Did you think about it? How could they determine weither a particular
wheel was tilted? Or I don't quite understand how they could setup all the parameters in 15 min before play?
Have you ever intercoursed with the Eudaemonians?
Thank you.

Sent: Friday, May 04, 2001
Subject: From Dimtry about roulette

It is Dimtry from [Moscau] again.
Let me ask you some questions.
I need to understand exactly what means 'touch down
time' in your data.
Is it the time when the ball at first time hit the
rotor after leaving the upper track of the stator and
reflecting from "diamonds" on the stator (is it the
first touch on the rotor)?
Or is it the time when the ball entered the winning
pocket and stopped there waiting for a dealer to stop
the rotating rotor to claim the wining number?

Did you time the touchdown moment by reacting to
sound of the ball hitting the rotor or how?

Further, Is the target slot a slot (pocket) onto
which the ball hits when it falls from upper track and
have been reflected from diamonds (that is it is the
place where the ball at first touches the rotor)?

Did you analyze the motion of the ball from the
time it left the upper track on the stator until it
crossed from the stator to the rotor?
Does this always take the same amount of time?
How may this time vary?

Sorry for so many questions:)

Also, if it is not difficult for you, could you send
me more data for analysis.

Thank you very much.

Sent: Friday, May 18, 2001
Subject: RE: From Dimtry about roulette

Every Russian casino has metal detectors and most of
them prefer to let in well-dressed rich people, so you
should not be surprised at that occasion.
ES - early surrender. It is much more player favorable
than Late surrender. But most casino manangers want to
have unexperienced players and they are afraid of pros
so if you win a large sum (even by luck) or constantly
win they will ban you not depending of either you are
a counter or not.
I am going to purchase a wheel but don't know what to
prefer: to buy a second hand wheel or buy a new wheel?
If it is indifferent than of course it is preferable
to buy the second hand one for less money.
As for the method proposed by your, Do you imagine how
you can realize simulation staying in casino (you
should use your hands for operating the simulation and
it is very disclosing)?
As for neural networks, we have some good cracked
software. A friend of mine used it for stock market. I
never paid much attention to that. If the physic fails
then I have opportunities to try networks.
I live in [Moscau] but periodically go to St. Petersburg to
play in local casinos.
Good Luck


KIM-1 computer
The KIM-1 (Keyboard
Input Monitor) micro -
computer is the grand -
daddy of all 6502 micro -
processors systems.
It was originally created
by MOS Technology, ...

Sent: Monday, August 06, 2001
Subject: Re: have you seen this?

I have looked at the roulette. I think it might be easier to conduct data taking with less heat on that roulette especially if it is located in a small gaming hall (which is usually stuffed with slots and vide poker machines). Last weekend I spent doing video experiments with my wheel( my wife's starting to boil:) because the room is getting like a laboratory). How did you describe the ball trajectory after the fall-off moment in your algorithm? I found that after falling off the side wall of the stator (where it is orbiting) the ball always hit first one of the four vertical diamonds and never hit first the horizontal ones. That is, after being deflected from the first vertical diamond encountered, the ball may hit the next horizontal diamond or hit the side wall of the stator or go right to the rotor or anything else he likes :). Was it true with your wheel? I ask this because I want to clear if this phenomena is true for all roulettes.

If you don't remember maybe NOWSCAPE-2 can recall it.
I am now fighting the most difficult part of the ball trajectory: the path from the fall-off moment when it
loses contact with the side wall of the stator until it hits the rotor.

I'd like you to answer these questions very much!
Thank you

Pondering the curves - analysis on the Apple II.

 Correlation between REVolution time and POSition.

Sent: Monday, August 06
Importance: High

Dimitry --
What you describe is very similar to our observation. I cannot tell you 100% exactly now, but your description seems correct. I think that the following may be true: Whether the ball reflects from diamonds or not, almost zero energy is lost, in most of the collision and the speed of motion of the ball toward the center will be roughly the same (although it will be offset somewhat in the tangential direction). The way I remember it, is that we ignored the ending sequences of ball motion. We did not study it as carefully as you did. I would be interested in seeing some pictures, if you have them and, with your permission, of course, I'd put them on the web site. But perhaps it is too early for you to give up your hard earned secrets. :) We thought that given the knowledge of where the ball WOULD land, if there are no deflectors (or separators between pockets), that would be enough. I still think so.

How are you timing the ball and wheel? Manually? What kind of computer do you plan to use in actual play? What computer language do you use to program the machine? Have you considered simulating (via training) the motion of the ball and wheel, instead of brute-force predicting the outcome via formulas involving second-degree polynomials? (We did not do this.) Perhaps you are using a neural net. I'm sorry, I forgot what you said, and my previous email from you is on my old computer -- I have not yet migrated old email to this new machine.

I recently got roulette related email correspondence from a silly person in Slovenia, if you want to see it, I'll send it. Keep a spot for me In Moscow -- I may still be able to come.

To your continued success,

Sent: Wednesday, September 12, 2001
Subject: I wish it were just nightmare

First I saw s&p500 futures tumbling down below the
monitor screen. A few minutes of jokes about poor
profits changed to a state as if we were watching a
film-catastrophe when the news began to come. Everyone
was shocked. A taxi driver who drove me home was
cursing out the Arabs as russian people know the
problem very closely. Those buildings blown up in 1999
were just 10 kilometers from our house.
I hope no one of your friends and relatives have not
Russia is really grieving.
US must forget the human values in relation to
terrorist states as they forget them in relation to
the civilized world.
Write how you are.

Sent: Thursday, March 07, 2002
Subject: RE: approximately 1 in 18


Well you have got very similar results to me, nice to know that someone else has done it too.. Before me I might had.

Well thought I would tell you that the Psion series 3A has speech synth built in. It also has a dedicated free running clock chip that you can access through your program. This can transmit directly to your tele-coil hearing aid!

I obtain approximately 1 in 18 on any wheel instead of 1 in 37 (European). I can double my money in 2 hours. Wondered how well you do??

Nice to share some ideas with someone with similar interests!

Contact me on
Mark Howe

Date: Thu, 7 Mar 2002
Subject: approximately 1 in 18

Hi, Mark -- It's interesting to know you succeeded. We did not peruse
our machine further, we needed refinements in timing. In what section
of the world do you live? I' curious about what it means when you say I
obtain approximately 1 in 18 on any wheel instead of 1 in 37 (European).

Sent: Thursday, March 07, 2002
Subject: Re: approximately 1 in 18


I live in the UK England!

American roulette is a little harder with the two zeros!

Well when I found your site, specially with the graphs, WOW, just so much
like mine, uncanny, even the ripple effect of where the ball hits the rotor
and finnishes up!. TO only a few people, that understands what you have
really achieved! Congratulations!

Well could you, or did you place 1 chip on at a time, that is what I do!
I hit the one number every 18 spins or so on average, yet to chance I should
only get 1 in 37, and can do this all day long. I made my program flexible
so that I could go into setting the equations up after any spin. This way if
drift occurred due to air pressure or temp etc I could instantly change the
program to suit. Unfortunately I never read the Newtonian casino book, or
came across anyone else who had done it, i thought i was the first, should
know better as I invent new products in my spare time and quite often come
up with ideas that have been done else where, well great minds think alike!

Down to successful, the hardest part for me was learning to program from
scratch, the maths was not to difficult!. Also I have been building radios
from the age of 5 years, so the electronics was not to bad either. Had
problems that I gather you will have had using the computer you chose at the
time, might be wrong though, well for 6 years after I had done the maths
using a Psion series 3 (not a 3A), i use to get fluctuations in the results
sometimes, bad enough to throw it completely out!
Then I bumped into a electronics whizz who had worked for Psion computers.
He told me that using the clock in a computer for real time applications ,
especially the palmtop I was using, was no good. This is because the timer
chip is used elsewhere in the computer, like to update the clock, if it has
one etc, run An auto switch-off feature, if you left the computer on. He
told me that I needed a Psion series 3A, this had a dedicated FRC timer
chip, it also had a speech recorder built in. I was on new ground, was not
long before I had a neat little system making me some money


Any more questions??

Sent: Thursday, March 07, 2002 1
Importance: High

Hi again --
We would have given our eye teeth, to have had a computer like the Piosn Series 3A :) As it turned out, even though we had to add the speech synthesizer and the infrared link, it was not a big problem, in retrospect, although at the time we did a bit of head-scratching.

I'm glad you enjoyed reading the roulette web pages; no doubt you saw the pictures also. So you know we made it a 2-man job, to do the actual casino playing. We, like you, placed only one chip per bet, as it makes no sense to bet anywhere else other than the predicted number. We also tapped into the clock chip of the computer to create our timer.

Thanks for the congratulations; I think you deserve more than we, because apparently you succeeded. There are quite a few people who are doing this -- I have had contact with several, especially Eastern block countries and Russia. One fellow in [] seems to be quite advanced. Many don't know what they are doing, I think. Estimating in which half of the wheel the ball would come down was my goal, but we stopped working too soon.

We could not adjust our computer on the fly, as you apparently can. We gave some consideration to building a machine with learning capabilities, but we chose to go the straight physics way, considering the decay of the ball and of the rotor as second degree polynomials. This you see on the graphs on the web page. What we did, was to take data from the wheel before play. Then we analyzed it to determine the parameters for our formulas. The we suited up and went downstairs to play.

You make no mention of a partner, for playing. Have you solved the problem of how to time the ball AND make bets? I'd be interested to know how you do the timing. Nowadays, I think it's possible to do the timing with a tiny TV camera, using fiber optics. We experimented with a simple device built into eye glasses, which would sense the rotational speed of the ball and rotor. We did not develop this further. Instead, we used manual switch presses.

Gang arrested after £1.3m win at Ritz casino

Three people have been arrested on suspicion of using James Bond-style electronic trickery to win almost £1.3 million at a London casino.
The Metropolitan Police said today that two men and a woman, all from Eastern Europe, had been arrested for allegedly obtaining money by deception.
They are accused of using a scanner inserted into a mobile phone and hooked up to a microcomputer to calculate where a roulette ball would probably fall after the third turn of the wheel.

The Daily Mirror newspaper reported that the trio attracted suspicion when they won the huge sum at roulette last week at the casino in the Ritz Hotel. They were paid £300,000 in cash and more than £900,000 by cheque as part of their winnings.

A Scotland Yard spokesman said "a significant amount of cash had been seized in the investigation".
In the game of roulette bets are placed on alternate red and black numbers on a wheel, numbered 0 to 36, and odds are paid out according to where the roulette ball lands when the wheel is spun.
The newspaper said that police were investigating whether they had used the scanner to calculate the speed of the ball when released and its probable finishing point on the wheel.
By the time the calculation were made the gambler would still have several seconds before the croupier called "no more bets".
Bosses at the exclusive casino, frequented by Arab princes and international playboys, became suspicious of the amazing run of luck on March 16.
All tables are monitored using video cameras and experts looked at the tapes before calling in Scotland Yard.

Two Serbian men, aged 38 and 33, and a 32-year-old Hungarian woman were arrested and police also seized their winnings and mobile phones.
The suspects were taken to a police station and then bailed until March 30.

Dr Mark Griffiths, Professor of Gambling Studies at Nottingham Trent University, said: "I haven’t heard of a laser scanner being used before but roulette mathematical systems have been used for years based on the speed of the ball and where the croupier puts the ball.

"It is reliant on calculations being done very quickly and when you have got technology involved it improves the chances.
"But the industry is well aware there may be underhand methods operating and when something happens once, word travels very quickly."
The alleged scam happened at one of the world’s most high-profile gaming houses.
The secret rooms of the Ritz Club casino are tucked away off London’s Piccadilly in the former ballroom of The Ritz Hotel.
It opened as a casino in 1978 and is now one of the world’s most exclusive, privately owned members-only clubs featuring chandeliers, ornate ceilings and plush red carpets.
A spokeswoman for the casino said today: "We have no comment to make."


From: George
Been looking at your URL [Pictures_desc_1984.htm] - do you still have the high profile double-0 A/R?

From: George
At 04:47 AM 10/13/2007, you wrote:
Subject: Re: : Roulette Wheel

Thank you for your prompt response; much appreciated. In the early 1980's following the 'incident' at Northampton I was asked to "find out" how it was done and to "do something about it". Following six weeks of an intensive research programme I designed the first low profile roulette wheel (Huxley Mk.3). I have been tracking most of the 'prediction' activity and have been endeavouring to stay one step ahead - my last design was the 'Starburst' wheel with triangular pockets that sends the ball backwards at low speeds.

I have been writing a book for the past ....twenty years; it took time to research every aspect, write programs etc. I have now retired and have more time in my hands. Referring to the double-0 A/R (what make is it?) I can envisage the following options:

(a) Purchase – please let me know what you have in mind;
(b) Exchange / Trade ? with a single-0 near perfect condition [low profile] Huxley Mk.5 wheel

I would appreciate the following:
(a) One or two photographs of the wheel; the photographs to be taken vertically i.e. the viewer will see the turret in the centre of the photograph. The photograph to include the plinth i.e the complete view of the wheel.

(b) A short video clip. Say 2 or 3 spins only. If you don’t have an Ivorene ball, a PTFE (or Teflon) or even a compound will be fine. I am afraid that the video should be taken with a ‘steady hand’ (or even a tripod), again, vertically to the turret so that the turret is in the centre. The video file may be *.avi or *.mpg.

I am willing to pay for both the photographs and video clip, including any DHL fee / costs. Please let me know whatever the cost might be. Also willing to pay in advance.

I have several single-0 wheels but no double zero, especially a high profile, and I would need to test the software with a high profile, different operating conditions, different balls and different inclinations (i.e. tilts).

I have now completed the ‘3rd’ generation software (written in Visual Basic 2005); some of the features include the following:
Video module (one of five modules)
· Identification of the winning number by means of a video camera / video file;
· Projection of the winning number at least 6 seconds before the ball exits the ball track (accuracy: straight up 8 out of 10 for horizontally true wheels);
· Projection of the winning number after ‘2-clicks’ for both wheel and ball;
· Deceleration / Service graphs
· And many more.

I attach one of my publications that you might find of interest. If you would like to read more on the first profile wheel, poker chips etc. please let me know.

At 05:23 AM 10/15/2007, you wrote:
hey is this projekt for real? and what was the result? I guess u weren't able to calculate the sector were the ball was landing. I just wonder how u clocked the wheel and the ball and the friction. Do u have a formula? im doing a similar projekt and i don't think its impossible to calculate the balls landing sector. Hope we can exchange some info. Thanks in advance

Dear Rickard --
This project was a long tome ago. The formula we used was that for a second degree polynomial. We timed the revolution and used those data as inputs to a curve-fitting program to give us the coefficients, a b and c for the plynomial ax2 + bx + c = 0. We used this formula for both the ball and the rotor. In practice, the rotor decay was almost zero, so its graph looked like a straight line. We clocked the ball and rotor manually.

2004.  Interesting: The Hungarian blonde, the two Serbs, the hidden gadget … and the very angry casino

The group was arrested, their mobile phones seized and hotel rooms searched. Although a lot of cash was found, the three were released on bail because technically they had not broken any laws. They could not be named for legal reasons.

“The operating principle of the gadget is relatively simple,” he said. “A laser beam measures one position of the inert ball, say, on zero, then when the ball is spun it measures its speed, tracks its movement and works out its likely landing spot.” (HA!  -- NOWSCAPE Editor :)

Google LINKS

From: pieter@***-***.be
Subject: === Roulette ===
Date: Wed, 31 Dec 2008 05:17:15 +0100

I’m interested if you can send me the raw data of your roulette of around 100 spins.
What I’m gonna do with it?

I have written a Visual Basic (or I am working on it… because at this moment it’s not finished) program where I can track/record or enter following variables:

- The speed of the rotor (the program calculates it with a built in chronometer; I have to click stop and start after each spin).
- The speed of the ball (program can calculates it too with a chronometer when you click stop and start)
- The diameter of the wheel

Now I wanna see if there is a relation in this data… Finally I want to make a formula to predict the landing zone of the ball.

Finally, If you can predict 4 numbers where the ball will NOT gonna land, you can beat the system (just bet on all the numbers except the 4), you will lose a lot of money, but every time will win more than you lose…

Do you have tips for this? What do you think about it.

If this works, I can write a program in Java so I can run the software on a Nokia mobile phone. You can connect the phone with a Bluetooth invisible headset so you can hear the prediction.

Thx in advantage!

To:  pieter@***-***.be
Subject: Re: === Roulette ===
From: rich***
Date: Wed, 31 Dec 2008 11:40:15 -0500

Hello Pieter,

Yes, what you said is exactly what we did:

  1. Compute the speed of the wheel at a certain reference point.
  2. Compute the speed of the ball as it passes a reference point.
  3. Plot the intersection.

Here are a few things we learned along the way:

  1. The wheels are not always level or accurate. Some are tilted, or are cut at different angles. Each wheel is slightly different.
  2. The balls have different size, mass, and material. The ball is changed sometimes.
  3. Altitude and barometric pressure have an effect, but we didn't determine whether that was significant.

Be careful, the casinos are looking for these sorts of things. If you begin to win a lot they will look at you very closely.

Where are you located? (Belgium??)

Good luck!


Hi Richard,
First at all a Happy New Year!
Do you have already a formula to predict it (based on the variables I said)?
I have read a formula on your website (in the mails you published)… Does this one work?
I understand that there are things you can't calculate (like ball mass, material, etc), but I think (am I righ t?) that it's possible to create a kind of  "calibration" modus in the software, so that it can correct itself….
Normally the house edge on European wheels is only 2.37% (on one number they pay 35x your bet, and there are 37 numbers…), so if you can write software that can turn 3% in your advantage, you will always be the winner ofcourse.
I'am located in Belgium, yeah.
The casino security in Belgium is not very strong (but there is!), only high resolution camera's are there. And if I can create the software, I would travel around the world to earn my money J. I also think in Belgium it is not forbidden to beat it with this way, they can only escape you and place you on a blacklist of the casino…

To:  pieter@***-***.be
Subject: Re: === Roulette ===
From: richard***

Date: Thu, 01 Jan 2009 13:48:17 -0500

Happy New Year to you, too!

We found that the wheel was easy to predict. It slows down at a very steady rate. If my memory is correct, the rate is about 10ms per revolution. For example, the first revolution might take 1000ms. Then the next takes 990, then 980, etc. We timed the wheel as it passed a fixed spot, such as the 12:00 position. We tested the accuracy of our hand-made presses against presses we measured with our own wheel. We were able to make quite accurate measurements of the wheel and predict the future location fairly accurately several seconds into the future.

The ball is different. As mentioned, there are different slopes of the wheel, different individual wheel defects, different balls used, etc. Also there are the diamonds when the ball falls from the rim. These are not fatal problems, but they make it difficult.

We measured the ball at the point it passed the same marker we used for the wheel. We used the 12:00 position on the wheel as our marker. And, just as you said, we used the same chronometer (a computer program written in machine code that counted in milliseconds).

The wheel data, when plotted, make a curve that is a second-degree polynomial. AX**2 + BX + C. The ball also makes a similar curve with the same equation. The data need to be put into a curve-fitting routine to best fit the data. From that, you can predict the outcomes for a wheel, using that ball.

Yes, it should be possible to adapt to another wheel without having to collect additional data. But, I think that would be very difficult. My suggestion would be to get it to work in one case, and then think about changing it after it is proven. We learned a lot about the project as it progressed, and we found unexpected problems, etc. I'd suggest keeping it as simple as possible, then making it more complicated if you are successful.

I have been to Brussels. Beautiful city, wonderful architecture, excellent food, great museums. I visited the World's Fair site, the Gran Place, etc. My father was there in 1944-45, really enjoyed it, and always wanted to go back, but never did.

Don't underestimate the difficulty of your project. It will take a lot of time and be much more difficult than expected, with no guarantee of success. But, it was a lot of fun.

Please let us know how it is progressing. Good luck!