, ,

These numbers are a random sampling of 300 random digits from a book consisting of little other than random digits. The story of their development is an important chapter in the history of science and digital technology.

This excerpt is from the 1955 edition of the RAND Corporation’s  A Million Random Digits with 100,000 Normal Deviates

… it was an important 20th century work in the field of statistics … produced first in 1947 by an electronic simulation of a roulette wheel attached to a computer. The results were filtered and tested to generate the table ….

The RAND table was an important breakthrough in delivering relatively robust random numbers, and had no precedent. The contents were available as a book or on punch cards. The table’s main use was in statistics and experimental design of scientific projects, and especially useful for Monte Carlo simulation.

The Monte Carlo method of simulation, driven by repeated cycles of a process based on randomly-generated numbers, pseudo-random actually, more on that in a moment [1] is widely used in cryptography and operations research today. Without a reliable source of random numbers, such simulations had been impossible prior to the RAND table, despite the relative simplicity of the underlying concept.    

Thankfully, A Million Random Digits… was rendered obsolete by the development of high speed computers in the 1960’s. Comparable or superior quality random numbers could be generated by an IBM mainframe running SAS in the 1970’s, and with MS Excel on a PC (running Windows with a 166 mHz processor!) by the late 1980’s.

The book was reissued in 2001 (ISBN 0-8330-3047-7) with a new foreword by RAND Executive VP Michael D. Rich. Both the digits and their deviates are available for free online, as: A Million Random Digits and 100,000 Normal Deviates.

The reissued version is available from Amazon.com, and includes a long string of very funny user comments and reviews.

[1] That is my facsimile of a footnote. I wrote a post about the the relationship between pseudo-random number quality and strength of encryption e.g. using XOR and other methods, for random numbers generated in Java. At the end, in the comments section, there was a bit of discussion about the Mersenett Twister algorithm and sample results for some rather elderly PC’s generating pseudo-random numbers.