I have been playing VB for a while and getting some progress in the past week. I recently read the book called how to visually detect biased roulette wheel and had been able to spot some defects on the wheel. The defect on the wheel I am playing is wheel head wobble and cause a region to appear more frequent than normal. The ball also land 45 degree to the rotor after it hits the diamond so there is a good chance to test the biased region. My way of playing is to actually bet the biased region 1 unit each first and then when Visual Ballistic indicates that the prediction region is near the biased number region, I would bet 5 units. However, if the 2 regions are different, I would leave the biased number bet as it is and only bet 1 unit for the VB prediction.
It is too good to be true that I have achieved edges of 100percent using the above method. Has anyone used a similar approach? I also realized that different wheel speed has a different effect on the bias as a ball strike the rotor faster would likely to mean that it would bounce further away to the defect than if the wheel speed is slower. However, it would cause suspicion if you only bet after the dealer spin the wheel or the dealer would not spin the wheel at all if you are playing heads up. Therefore, I would use an average wheel speed for each dealer and get data for both directions. The dealer has muscle memory and would not be too far off the average.