contin tests the equal distribution of a random fraction from zero to one. This type of random number is continuous, because it is a real number.
Thirty tests are run for 1000 random numbers from zero to one. The numbers are sorted and then compared to a smooth distribution of 1000 numbers from zero to one. The differences between the random numbers and the population are analyzed with the Kolmogorov- Smirnov test. The 30 K+ numbers from the 30 tests are then applied to a second Kolmogorov-Smirnov test and the results printed at the end. Likewise, 30 K- numbers are run through a third Kolmogorov-Smirnov test and the results printed at the end.
The chi-square for each of the thirty tests has 29 degrees of freedom. The global chi-square test also has 29 degrees of freedom for the same 30 thousand numbers.
The f statistic is a one way analysis of variance for the means of each of the 30 tests. If this test passes, then the means of the thirty tests differ because they are random. The test is rejected if the f statistic exceeds 1.45.
The confidence levels for these tests are 95 percent.
contin doesn't need parameters. The scope of the test is fixed, with 30 tests of 1000 numbers.