Should the numbers for the experimental probability be close to the theoretical probability?

Not necessarily. There may not even be a way to work out a theoretical probability. Furthermore, there is always a chance, however small, that the experimental probability is way off.