You're starting to get into the concept of probability distributions. What you've been mucking around in is the uniform distribution where every event has an equal chance of occuring.

I think the easiest way is to define each event having a Range of values. If you have two events where one event is 90% and the other 10%, then you can get a random number between 0 and 9 and then the numbers between 0 and 8 are event one, and 9 is event 2. This can be generalized for larger numbers of events.

Of course, if you use modulo then your distribution will in fact be different because, say, rand() % 10 is NOT evenly distributed. That is, some numbers will occur more often than others. This will end up skewing the ranges given above.

Hope this helps.

edit: I should probably mention solutions around using % for getting your random numbers instead of just saying it will cause problems. This site seems pretty good at explaining techniques to get a better distribution.