(Ab)normality calculator.

I am looking for a web app or else a windows app that implements a theory I have for calculating normality or abnormality (evenness or unevenness of distribution of decimal-place digits) in numbers. The app would accept three inputs: 1. a numerator (which, depending on how advanced the software, may be the product of an inputted calculation or else it has to be a literal value) which is to have its (ab)normality measured 2. the number of denominators it runs tests with (which gets multiplied by two, to deviate positively and negatively from the numerator in equal amounts). 3. The precision (number of decimal places) that the calculations use. The higher the number, the more precision. 4. (optional) choice of base number system, e.g. binary, decimal, etc. Let's say the numerator is 592 and I choose ten denominators (a span of 20, between 582 and 602). It would start with the first number that is 1 greater than the numerator, and divide the numerator with that denominator. I.e. 592/593. It would calculate the result to an ideally large number (user choice, let's say, 20,000) of decimal places, such that it can achieve a distribution of numbers that gravitates (or is expected to gravitate) toward a mean distribution. It would then measure the distribution by counting the frequency of occurrence of each number. It would then make a calculation that is the frequency of occurrence of each of the 10 possible digits, divided by the number of decimal places (in this example, 20,000). In a normal number computation like pi, each digit occurs with a near-to-10% chance for each of the digits in base 10. Second-to-last, it subtracts the most-occurring-number frequency (%) from the least occurring frequency. In the case of 592/593, 6, the most occurring digit, occurs with a 10.1% chance, and 5, the least occurring digit, occurs with a 9.9% chance. Subtract 10.1% from 9.9% and you get 0.2%. This 0.2% is called the products' abnormality. In this case we chose 10 denominators (we could choose a thousand, or a million, however many tests we wish to run!). So 592/594 is next, and then 592/595 and so on. When it reaches 592/602 it goes on to do the less-than-numerator denominators: starting with 592/591 and then ending with 592/582. At the end of this we will have a set of 20 measurements of abnormality of the product. Final step: we calculate the mean average of all the (ab)normality results for the given numerator. The HOPE is that this metric is how 'pattern-rich/fertile' said numerator is. The more tests you run, the more precise this result will be. EDIT: I wish for three outputs: numerator greater than denominator calculations denominator greater than numerator calculations mean average of all abnormality results.