Since I have spent years in roulette research and have a connection with people who have same interests and who achieved the most of this game, I think writing article about roulette computers that average Joe can understand is more useful than empty promises found at some sites.

It’s been six months since I’ve started a new project of converting current FFZ roulette computer to make it a talking one, with the new name “the FFA”.

Adding audio and making it talk took me less than a day but remaining time I have spent in rebuilding close to 3,000 lines of the program to optimise it for the best performance.

The new unit uses machine language, the fastest, the best and probably the only one suitable for such application.


What are the real reasons for converting the FFZ to FFA?

I still consider the FFZ as an ingenious and unique design most suitable for a skilful single player. There are few reasons which pushed me towards FFA’s direction.

One reason is players who already own the FFZ and has requested of me to build it.

Some were telling me “the FFZ is great, extremely accurate device but I want to play more, and I need it to be made for team play”.

Some users found “zap” stressful and prefer talking roulette computer. The others were saying, “keep the zap” because they already tried roulette computers which use audio and hearing aid earpieces from some other sellers and found it useless. The question was can I make my system to talk and still have the same accuracy as the FFZ has.

I have two diplomas, telecommunications and electronics and roulette is my hobby which sometimes becomes an obsession. It took me many hours, but the answer is yes, and it is done on time as it was scheduled for completion for the middle of June 2008. Not only that system is converted to use audio instead of zap, but the software is completely rewritten.

To make everybody happy I decided to make it both, new updated FFZ with the latest software and the FFA as a new system.

The reason for reconstruction of all programs is to improve systems flexibility, accuracy, additional options etc. other words to make the system unlimited.

If you have been following this site, you could find that Michael Barnett (“Casino Consultant”) has purchased one of my FFZ units for his testing. Regardless what some competitors are saying about his test, the fact is that he was never interested in buying and testing their units but only mine. It is because anybody with some roulette understanding knows that others are running sites to scam people and that isn’t a threat to casinos. I did write more at my forum site about this case, and here I will give just a brief overview.

His findings were that FFZ predicts correctly on wheels with common drop point but on the levelled wheel, he couldn’t get enough predictions. There is a gossip that I deliberately supplied him with a faulty roulette computer, but that’s not the truth. I provided him with an excellent working unit as I would supply to any other purchaser. His testing ended with a dispute; he refused to return the unit for checking because he did some modifications on it. Also, he did not want to provide any additional information about the test so I could analyse why he couldn’t get enough predictions. Every single person who owns the FFZ told him that they are getting about 80% predicted spins, but the casino consultant insisted on his 20%.

As the only logical answer I could think off without having data analysed, is that the FFZ was tested on a single wheel which is out of FFA’s specified parameters. Casino Consultant wants to purchase latest FFA system as well, for his research. However, I refuse to sell new system except to people that I know and can trust. What happened is that during redesign I also revolutionised prediction on standard drop point wheels. Just during the testing period, I managed to profit more than the total system sales, so my decision is very justifiable. Of course, there are other reasons for such a decision.

Anyway, some of the Casino Consultant’s comments were another reason why I decided to rebuild all programs. Since from his test I didn’t get anything I expected, I don’t have any interest for FFA to be further tested by him.

I believe my expectation from the test outcome was standard and that would include results from different wheels where different kind of ball was used. But I received nothing of that. I didn’t make changes for the Casino Consultant but to make the system most accurate most practical and the most flexible for any situation. Since his alternative option was to purchase it under a different name all I could do is stop selling it. My decision was highly welcomed by other FF system users.

The FFZ would predict where the difference in between ball rotations at the time of prediction is in between 120-255ms and the most of casino wheels are within that range.

New changes make it more comprehensive, and for the FFA system, it is from 70-392ms. Such wide range allows the system to predict any wheel. (Here I didn’t say that and advantage is possible at any roulette wheel). Recently I have tested it on one of latest new wheels with very low ball deceleration, and the FFA was predicting without any problems. How the system will perform on all types of wheels, I only don’t know until I get more feedback from people who will use it.

With this article, I will avoid writing about other roulette computers scam especially the ones built inside mobile phones which are entirely incapable of timing applications. Also, mobile phone uses an audio player who takes 1.2 to 1.5 seconds to activate. Such time losses aren’t affordable for serious roulette prediction. You can confirm this with any skilful software developer, and more information is available at my forum.

Therefore instead of talking about something that doesn’t work I will try to explain what the roulette computer is.

Since there is nothing to compare the FFA too, I decided to compare it with FFZ and to explain what can be expected from roulette computer, what is essential, how much it is important and why.


Roulette computer accuracy

For roulette prediction, this is the most critical requirement. Without this anything else is irrelevant. For roulette computer to have needed accuracy few factors are required.

Hardware and software that can accurately measure timings and to process all calculations in short time

Both systems FFZ and the FFA are using the same microcontroller programmed in machine language. Hardware is set to measure time with 0.000001 seconds accuracy; it’s far more than enough. So there is no real difference here. The software is capable of making more than 100 needed mathematical calculations during a single spin in less than 10 ms.

Both systems are built based on same procedure; both use right hardware as it is required therefore if I mark it from 1 to 10 I would give 10 for both systems. On this scale, if a mobile phone is used it would be 1 and PDA would be 2.

Ball clocking errors correction

It is the first feature I’ve done right, without that there wouldn’t be much progress in any further development.

Both systems use similar methods. FFZ is slightly more accurate in the ball timing and rotor calculation than the FFA. Smart system design can avoid some parts of the rotor calculation, which FFA must have. With the FFZ we can clock the ball entirely at a different position then rotor. However, since the FFZ displays a result, an additional calculation is required. Therefore the system can create some errors. Also during the play, the user is dependant on his observation which can produce further small mistakes.

Time from the last switch clicks until we get information is in average 0.5 sec.

On the other side, the FFA pronounces number which takes about 0.7 sec, but it is easier and more accurate information then observation. Because it doesn’t use visual indication, it is more accurate. However, the FFA as every talking roulette computer must calculate rotor position from the time we clock it, which can inject more errors if compared to FFZ. It is logical, for more time we calculate the error may increase.

When I put it all together on the same scale, I would mark FFZ with 7 and FFA with 7. But only, if the FFA has enough time to wait until rotor comes in a particular position.

Mobile phones and PDA’s, because of the already wrong measured time just cannot perform any of these tasks. Also, they take too long to get information out. As usual, sellers of such devices would suggest doing ball clocking earlier which makes things worst. Timing the ball when it is only 0.5-0.6 sec per rotation and where the difference in between rotations may be just 40 ms to get prediction is wrong. In such case hardly that any accurate prediction would be possible even if right hardware as FF is used.

How roulette calculation is made and what does it include?

Some computers use the camera to record a casino wheel; then the player needs to go home, play video in slow motion read ball timing, approximate it because video frames are distanced 40 ms and enter data into the computer. While they complete all of that and come back, probably the next day, all they can see is that everything has changed (if they see it, that is).

When I come up with IQ solution where the system is capable of learning ball deceleration curve within few milliseconds accuracy, all they could do is to start claiming the same. It doesn’t make sense if the computer can learn to the point of few ms accuracies then why to use a camera to get necessary parameters with 40 ms jumps. If such data has to be manually added, then it is evident that computer doesn’t use any complicated algorithms to define ball parameters.

First of all roulette prediction is dynamic – taking timings as explained or building graphs are pointless. What is valid now after only one hour may be completely different, or the next time you go to play. Spins next to each other may be different as well. It is such a naive way of thinking. But who cares if the designers objective is to sell the computer. If someone is so uninformed to buy it, after investment and failure, he will probably purchase and a camera, which is standard camera valued $100, but he will pay for that from the seller for about $800 – $2000.

Both of Myrulet systems automatically learn the wheel while we play.

Learned information is essential for proper calculation. It needs to be current because wheel parameters may change in short time.

The FFA learns wheel information for both ball direction independently. Therefore, the calculation for ball clock way and the anti-clock way is entirely independent.

roulette differenceA few weeks ago I monitored the FFA’s data during casino play in both directions for two days. Yes, there was a difference in between ball directions, the wheel wasn’t symmetrical, and parameters have changed from day one and day two.

The difference in between day one and day two in ball deceleration was 20 ms. ( it is not a real ball deceleration but only a gap in between rotations at a common point of prediction) It doesn’t sound as much, but it caused the ball to travel shorter by 2 seconds. That is a difference by close to 50 pockets. Now, imagine if someone used the camera to set the system then came next day to play.

I better cut to the point, so we don’t end up with a very lengthy article.

Merely the FFA is better, here it gets 8 out of 10, and the FFZ gets 7. Both systems handle some things well, but the FFA wins.

Time of roulette prediction

Some wheels may not be predictable at 12 seconds to the end, but a prediction about 8 sec may be very accurate. The main reason for this is phenomena of knee point where the ball suddenly may change deceleration. FF systems due to specific design can target particular time of prediction to take advantage of such wheel and to keep the constant advantage for all spins. All demos that I have seen from other people trying to predict roulette is based that player decides when he will clock the ball. Those results that time of prediction once can be 12 sec to the end next time 5 or 8 or 11 etc. They explain it as a good thing since the player may be suspicious placing bets always at the same time. It may be correct but what is the point if the prediction is not accurate.

If you continue reading you will soon understand how hard roulette wheel prediction is and why every single point is so important. Neither the FF system will have every spin predicted at 100% every time, but it will be at least in the targeted zone where an advantage is possible.

This scale is from 1-10, and both systems can get 8.

Additional roulette computer features

In this group comes offset adjustment and possibility of saving data.

The FFA can do it better, it can save few games at one time, and it can adjust offset independently for each ball direction. How much is that important?

For myself, there is not much I would need to adjust. If I know that the offset is +9 pockets and if system predicts number 34 I know where to place bets with an added offset in my head. However, after some time and few hours of play, it may be better to adjust it. Or perhaps if two or more players are using the system, it may be easier.

The option of saving the current settings to use it some other time, or to analyse the data. Well, I thought this is more important then what is. After serious research, I found out that it is not useful function since it is always better to let system learn the wheel from the start. Conditions on the wheel may change, setup data may be useless. Since it may be a change, the system may have a problem returning to now parameters. If we start from the start, it will take only 2-3 spins, and we can have fine predictions. Therefore saving data feature is useful for research or practising. From 1-5 FFA gets 5 and he FFZ gets 3

Practicability

The good side for FF systems it that they do not use a display for settings. All system settings as re-tuning and adjusting can be done at the table while we play.

Hardware set up for FFA is more complicated than the FFA because it needs more. FFA setting is more comfortable than the FFZ’s because of the system talks; with FFZ we need to count pulses.  For example, the FFA, while we play, can tell us everything about the wheel that we need to know. FFZ is practical when we do ball clocking because it can be at any position and utterly independent from rotor clocking. However, all of that isn’t much important.

FFA=4 FFZ=4

Roulette computer security

No system can be secure enough. In my opinion, the best is not to get noticed. Both systems can be locked in the case of emergency without even touching them. I fell more secure with FFZ; it doesn’t use radio communication.

FFZ=4 FFA=3

Roulette computer team play

The FFZ is not designed for team play, it can easily be modified if needed but the player would still need to see the wheel.

FFA uses radio communication and spy earpieces inserted into player’s ear channel.

Amount of players with the right set up is unlimited with an option that player can have nothing more than individual spy earpiece invisible. It means the player doesn’t need to have Bluetooth receiver or induction loop around his neck.

FFZ=2 FFA=5

If we add it all together and make a table we would get something like this:

FFZ FFA MAX.
Hardware 10 10 10
Error correction 7 8 10
Calculation 6 9 10
Time of prediction 5 5 5
Extras 3 5 5
Practicability 4 4 5
Security 4 3 5
Team play 2 5 5
TOTAL 41 49 55

 

What does it mean and is it good enough?

The answer is in next article.