Roulette Not Random!

Apparently, a successful roulette gambling system is possible.  It doesn't involve betting on red after black comes up ten times in a row, either.  Nor does it involve doubling your bet every time you lose.  The simplest such system involves merely timing how long it takes for the ball to make one or more revolutions of the wheel.  More complex systems would require a pocket computer and take into account the ball's mass, the friction between the ball and the track, the angle of the track, the rate of deceleration of the wheel, etc...  Of course, all of this proceeds on the questionable assumption that the croupier will spin the wheel smoothly so that the ball remains on the track as long as possible, as opposed to pinging off the slots.  The idea is that one can predict, often enough to tip the odds in your favor, in which third of the wheel the ball will fall.  And since you can bet on halfs or thirds of the wheel...  The New Scientist article is here.  Here is a puzzling question.  If it is possible to develop this sort of gambling system for something, is that something not random? If so, then what room do "laws of nature" leave for randomness?