SRL Archive - SRL Snippets

function RandomBinom(n, p: Extended): Extended;
var
  x: Extended;
  i: Integer;
begin
  x := 0;
  for i := 1 to Round(n) do
    if Random - p <= 0 then x := x + 1;
  Result := x
end;

Cheers,
Cynic.

I understand what this is and what it's doing, but what's the use case of this for srl?

And why make n an extended variable as opposed to an Integer or LongInt since it's rounded anyway?

And then side-question, what's the range on Random when it isn't passed an integer value?

Edit: Use case would make sense as a random number with a binomial distribution. Doesn't GaussRand() cover that though?

From: nivek1989
Edit: Use case would make sense as a random number with a binomial distribution. Doesn't GaussRand() cover that though?

Gaussian/Normal Distribution is continuous whereas a Binomial Distribution is discrete (only allowing for specified values like an integer range).
Allsooo… Binomial is used to 'count the number of successes until failures' in a line of independent success/fail experiments.
So, they have different applications.

From: Turpinator
Gaussian/Normal Distribution is continuous whereas a Binomial Distribution is discrete (only allowing for specified values like an integer range).
Allsooo… Binomial is used to 'count the number of successes until failures' in a line of independent success/fail experiments.
So, they have different applications.

This isn't counting the number of successes until failure however, this is counting the total number of successes over n times. Maybe I'm misunderstanding something, but this does not seem to break at the first failure, so many failures would likely occur over the course of this count? So what I do seem to gather is this is roughly the same as a Gaussian distribution, however focused on an Integer return, which could be done via Round(Gauss())?

From: nivek1989
So what I do seem to gather is this is roughly the same as a Gaussian distribution, however focused on an Integer return, which could be done via Round(Gauss())?

For "low" values it isn't even roughly Gaussian.