The importance of proper seeding of the Applesoft pseudorandom number generator

Abstract

The popularity of the use of Apple II/IIe hardware in medical applications is evidenced by the variety of articles devoted to the use of this hardware in biomedical research and medical practice. Many of these applications are of a stochastic nature, including simulations, and depend on the Applesoft random number generator. Unfortunately, the Applesoft generator has exhibited a number of flaws, including lack of uniformity of generated sequences of numbers, and serial correlation within such sequences. This paper provides evidence of these problems for a variety of seeds, and provides suggestions for acceptable seeds which ensure that the generated sequences of numbers pass two fundamental statistical tests for randomness.