Since you have no reply yet, I could give you my simple thoughts.
First, It would be interesting to know more about those automatic roulette tables. Can you see a name of the manufacturer or something? Are they inside a casino or in a pub or what? How much can one bet at once? Are they slot machines? How late is “no more bets” called? How does the wheel speed vary? Just any information like that, because machines like that offer an interesting potential. I have searched for an internet casino with live roulette wheel, so one could make extensive observations from home. But the only casinos with live roulette movies which I have found, require bets to be placed before the ball is spun. I suppose that this is not the case with your machines.
I would suggest a bit of programming. Maybe on a programmable phone or a pocket PC, or on a Basic Stamp for the ambitious. The computer should have a button for input and some form of simple output, either a vibration or a beep soound. The device could be in your pocket with your finger on the button.
Maybe Forester can offer you a nice solution by adapting one of his roulette computers?
When the ball is spun, you make one click on the computer button. Now the program starts measuring the time. In 25,5 seconds the ball will hit that diamond. Now you need to time the wheel speed.
When the zero passes that diamond, you make another klick on the button. (I say zero, because it is green and easier to spot more accurately, I suppose). The program notes this point in time. Next time the zero passes the same point, the diamon, you click again. Now the program knows how long time it took for the wheel to complete one revolution and hence the speed of the wheel.
Consider this example:
Time: Event:
0s Click for ball spun, game starts, 25.5 s to impact with diamond.
2s Click for zero making 1st pass by that diamond.
5s Click for zero making 2nd pass by that diamond.
Now there remains 20.5 seconds to impact.
Wheel speed is 5-2 = 3 seconds per revolution.
The wheel will make 20.5/3 = 6.83 revolutions until the ball hits that diamond. This means that the number one sixth (60 degrees) before the zero will be under that diamond at impact (0.83 = 5/6 after the zero, ewuals 1/6 before the zero). Say that is number 16 red.
The computer signals this to you by sounding a beep or making a vibration when number 16 passes by that diamond. It could make that signal for each revolution of the wheel. I.e. when you hear the beep, the number at the diamond will be under the diamond again when the ball hits that diamond.
This was a simple version described. Several adjustments could be done for better predictions. For example, you could continue to click the button each time the zero passes by that diamond. This would have several advantages:
- Human timing errors would be reduced by averaging multiple clicks.
- The later the wheel speed is measured, the less time will measurement errors have to grow.
- The program could take the decceleration of the wheel into account.
You will notice how the computer changes its prognosis as you update it with new clicks. If predictions are not fairly stable, you could avoid making a bet that game, because maybe some measurement went wrong.
Another thing to consider is that the scatter, the average distance which the ball will jump once it impacts with the wheel, will vary (increase I suppose) with the speed of the wheel. You have ovserved an average scatter of 1/4 of the wheel. It should be shorter when the wheel is slow, and longer when the wheel is fast.
The program could account for this because it knows the speed of the wheel. But it needs statistical inputs about how the scatter depends on the wheel speed. Then the computer should send a signal when the predicted number passes by that diamond, i.e. it should tell you directly which number to bet on (together with neighbouring numbers).
Further still, maybe it could be a good idea to time the speed of the ball more carefully by making clicks for its revolutions too. This makes things more complicated and starts looking alot like you need Foresters device more or less as it is.
Overall, I think that you have a nice opportunity with that tilted machine. You could even take advantage of machines without tilt, but a ball which hits the same diamond each time, that is very nice! It would be very interesting to hear how this works out for you.
EDIT:
Okey, I misunderstood the 1/4, it is not average scatter, but distance between diamond impacted and wheel impacted. So this 1/4 is independant of wheel speed. However, this is not important for the concept I’ve outlined above.
