I'd like to go back to the birthday problem, because I don't think it was answered well enough yet.
Instead of asking what are the chances of at least two people sharing a birthday, let's consider the oppposite situation, where no two people share the same birthday. Then, we can subtract that probability from one, and the result will be the answer we want (because it covers all the possibilities of people sharing birthdays).
Our first task is to count all the ways that the 23 people could have birthdays, where each person's birthday is different. Well, if we put them on a list, we have 365 possibilities for the first person. For each one of those possibilities, we would have 364 possibilities for the second person. We could make a list of all the possibilities (i.e. 1/1 & 1/2, 1/1 & 1/3, . . . 1/2 & 1/1, 1/2 & 1/3, . . .) but we can see pretty quickly that the total number would be 365*364. For each additional person, we multiply again by the remaining possibilities (365*364*363*362* . . . ), a process called a permutation. (the result is an incredibly huge number). This number we could divide by the total ways that the birthdays could be arranged (365 to the 23rd power) to get the probability that no one has the same birthday, which is about .4927, or 49.27%.
To find the chances that at least two people share a birthday, we subtract that probability from all possible outcomes, i.e. 1-.4927=.5073, or about 50%.