There is this idea in statistics called the "bootstrap." The idea is not crazy, but it's not quite right either. Works like this: data comes in, and usually some parameterized…