How's this for a thought:
One random result tells you absolutely nothing (unless it violates what you thought were the bounds). But as you increase your sample size, you can start to do Bayesian inference (people do this naturally without knowing the math behind it).
For example, let's say you hit a royal on your first hand. The attendant will come around to congratulate and pay you. Often they will ask you to "play it off" (play another hand) so that someone else can't claim it for another payoff. If you hit another royal on your second hand, the attendant will probably say something like: "wow, that's lucky, I haven't seen that in years." Now, if you hit another royal on your third hand, that's a "pattern" that will probably draw attention. You will probably notice a lot of "suits" coming to check you out, they might call a slot tech to do some checks, they will probably run a camera check and investigate the attendant, they may even call for a manufacturer's rep. They may even refuse to pay, claiming a malfunction, bringing in gaming for an investigation.
Three royals in a row is possible but it's a very unlikely "pattern".
http://en.wikipedia.org/wiki/Bayesian_inference
