This particular statistical problem has been well studied by mathematicians, so you don't have to simulate but can use a well known result. Or, if you must simulate (say, for homework), this can provide a way to test it.
When you have an event E that can happen 1/20=0.05 of the time, and you want to know a probability that the event E happens K (K=0,1,2,3,...,999,1000) times in N=1000 tests, the answer to that is the Binomial Distribution B(n=1000,p=0.05).
In Excel, this Binomial Distribution is available with BINOMDIST().
Example
BINOMDIST(49, 1000, 0.05, FALSE)
gives the probability of seeing K=49 events out of N=1000 tests when the event has prob p=0.05.
FALSE
is for one particular outcome of the 1000 tests, i.e. K=49.
TRUE
would give the cumulative probability of less than or equal, i.e. summing up the probabilities for K=0,1,2,3,...,49 events.