I’ve been testing on a slingshot wheel with the Hybrid and found a physical anomaly that reduces accuracy. It would affect any roulette computer that uses typical primordial charts (jump charts), including the Uber, by making the ball behavior and resulting drop time much more complicated to model. The anomaly is mostly dealt with by using the Hybrid’s equivalent of target. But with any setting that doesn’t use target, accuracy is reduced (eg, from possible 30% edge to 5%). The difference between actual and predicted ball drop time can be +/-2500ms (with predictions about 20s before ball fall) which is much larger than it should be. Depending on the individual wheel, there is still likely an edge, but not the maximum possible. Hybrid users can use the target feature, but this only deals with part of the problem, and you’ll need around 50% more spins than usual for analysis. I’m working on a solution that better models the ball behavior. This will be used in conjunction with target, then the hybrid will beat a larger proportion of slingshot wheels. I’ll work on it while I’m away and possibly the update will be available before I’m back.