Yeah it seems to be 9. Interrestingly though, one of the spots has only half the probability of the rest. This is because GE seeds a byte (0-255) and then checks in intervals of 30 for the first 8 spots. I.e:
Is seeded byte <= 30? Put FR in spot 1. Else continue. (Since this includes 0, this spot has a slightly higher probability than the other positions, 31 / 256 = 12.1%)
Is seeded byte <= 60? Put FR in spot 2. Else continue. (this is the best spot)
...
Is seeded byte <= 240? ...
Since a byte has a maximum value of 255, there only 241 - 255 left. So instead of an interval of 30 there's only 15 for the last spot. Fortunately though, this happens to be the worst possible spot :P
Conclusion: The probability for best FR is 30 / 256 = 11.71% (1 in 8.53)