You may have heard a roulette computer that gets predictions anytime in the spin by assuming the ball deceleration is linear is a great thing. After all, someone sells such a device so they publish claims to sell their product. You may have also seen videos of a computer for sale that gives the same prediction on each click as the ball slows down, and it may look impressive.
The facts about this approach, and its associated problems, are explained at nolinks://nolinks.roulettecomputers.com/earlypredictions.html (nolinks://nolinks.roulettecomputers.com/earlypredictions.html)
Also see a video demonstration at nolinks.roulettesystemanalysis.com/m/linear.wmv (nolinks://nolinks.roulettesystemanalysis.com/m/linear.wmv)
My computers have never assumed deceleration is linear, simply because it would be a simplistic and lazy shortcut with flaws. To be specific, assumption of linear deceleration would be acceptable if I only wanted to know approximately when the last ball revolution would occur on a wheel where the ball's deceleration was quite smooth. There would still be errors, but they would be managable on enough wheels. But if you intended to give a prediction based on something like which diamond will be hit, such a simplistic approach gives random accuracy.
Not many people will understand why, but essentially on a tilted wheel, you can have timings inaccurate to even 500ms (0.5s) and still get much the same prediction as another computer that is accurate to 0.000000000000001ms. But if you ask the computer to know which diamond will be hit, the smallest errors count, and you cant afford to take cheap shortcuts with the algorithm.
Anyway, anyone who has the tested device (FFA unlimited) can test for themselves. Anyone who has my devices can do similar testing. Then people can decide for themselves.
This is just one of many, many critical points that highlight the differences between computers. More is explained at nolinks.roulette-computers.com/comparison.htm (nolinks://nolinks.roulette-computers.com/comparison.htm)
Unfortunately telling the truth does have drawbacks.. like people that truth offends call you nasty names and lie about you, and some people actually believe it.
Just a quick note, forester still claims that his device can predict correctly anytime in the spin, but that for his acrobat, it is only relevant to know the difference in timing between revolutions. Here's what's wrong with this:
1. He contradicted himself. Clearly the accuracy doesn't exist with predictions anytime in the spin unless you only need sloppy accuracy for tilted wheel, no diamond targeting. So he says prediction anytime isnt required, but regardless it is still accurate with predictions anytime. That's not what the video testing shows, and its not what any of his customers will find if they repeat the testing.
2. Sure there is reasonable accuracy from a timing perspective if you look only at the difference between revolutions. But another problem is the ball deceleration rate is always changing, and his computers assume everything remains constant.
I'm not trying to discredit his product. I'm just explaining tests anyone can do for themselves.
A test for anyone with acrobat:
1. Get predictions at varying time in the spin (different ball speeds). You need it for 2+ diamonds, or accuracy is more based on full revolutions which means sloppy timings are ok (not a realistic test).
2. Do NOT click right to the end, or you flaw the test
3. Check the diamond prediction accuracy
Simple testing. Even just check iqe6 with predictions at varying times ffs, random. It is a glimpse of acrobat accuracy with predictions anytime.
bebediktus, keep digging.. do the testing for yourself. Do not rely on anyone's word - just simple testing and common sense.