Usage: ./{compiled program} [N] complexity] where N is the maximum integer to stop at, and complexity is the factor to stop calculation at (2, for the titular method.)
================================ Abstract from the Lawley paper: The Lawley primality test is a constant-time, probabilistic primality testing algorithm that determines if a number is a composite or a probable prime. Let n be the positive integer to be tested. If n is even, the number is composite. If n is odd, the number is "probably" prime; roll d6, and if you get 1 or 4*, return prime. Otherwise, return composite. The algorithm has achieved phenomenal experimental results: for the first 100,000 positive integers, it achieves a false negative rate as low as 6% and a false positive rate as low as 13%. These numbers place it very close to leading probabilistic primality testing algorithms (most of which achieve 0% error rates across the board) at a fraction of the complexity. In the event of a user leveraging the Banner revision, the program simulates a large die collection, each die being sized based on theorems by Hadamard and Poussin indicating that for large x the likelihood that a number within (1, x) is prime approaches 1/ln(x). For a given number, n, the likelihood that n is prime would be 1/ln(n+1). Obviously, since one cannot have fractional sides on one's dice, the simulated dice collector opts to use a die with a number of faces equivalent to the ceiling of the natural log. Therefore, n is "probably" prime if, upon rolling a d*, where * = ceil(1/ln(n+1)), the result is a 1. Rigorous tests have not yet been performed to determine the algorithms accuracy, but, when approached for comment, Poussin and Hadamard said the work "seems to check out."
*modified to 1 by the now-famous "Donovan Revision". *modified to 2 by the now-famous "Banner Revision".