Just to clarify, mine is an observation that is entirely possible to be false as I haven't verified. I've just hit it enough times that I *think* there is a possibility i.e. an unproven and untested hypothesis. It's still possible that I'm quite lucky or since I don't track it that I'm mistaken about the true number of times I hit it.
From another poster, the easiest solution from a programming standpoint is to have it determined randomly. This does make the most sense.
I'm still curious about why I tend to get 3 so often though. Although I have to admit it's not 100% reliable so if my theory that it's timing or somehow linked to the state of the lights could be explained by both human error or that my theory is false.