Random number generator
A random number generator picks numbers at chance from a distribution.
The most common type is a uniform generator, where each number has an equal chance. If a uniform generator yields decimal digits from 0 to 9, then each digit has a 10% chance. The other type is a weighted or biased generator, where the chances are not equal. A biased generator that yields 0 with 70% chance, yields 1 with 30% chance, would probably yield more zeros than ones.
True random numbers are impossible to predict. Many programs use pseudorandom numbers, which are not as good. A pseudorandom generator uses a formula and a state to calculate a sequence of numbers. Anyone who knows the formula and the state can predict those numbers.
Pseudorandom numbers can also have statistical problems. True random numbers are independent. A pseudorandom sequence may contain patterns where the values of some numbers change the probabilities of other numbers.
See also Pick_random_element.
- Wikipedia's article about Random number generation
- NIST's statistical tests for random numbers