Graph 1. I have also "crunched" the numbers of Bomb films using the "Smoothing Spline Fit" technique with a "goodness of fit" of roughly 0.84 (i.e., a fairly accurate generalization of the distribution of Bomb films).
 
 
This provides us with a way of visually understanding general patterns in the data. My colleague at Hiroshima University, Takao Shõhõji, an internationally respected scholar in computational mathematics, interprets this graph as showing periods of stable increase and periods of stable decrease, with some plateaus.
 
 
However, we both agree that, without a very elaborate statistical model that compares many different factors, we can say little else with certainty. (Atomic Bomb Cinema, 171-72).
 
 
 
Graph 2. By *"normalizing" the numbers of Bomb films to the numbers of all films, and comparing them, I also found, statistically speaking, that during slumps in the film industry Bomb films often outperformed all films (Atomic Bomb Cinema, 171-72).
 

* The number of bomb films is multiplied by a factor that is derived by dividing the number of all films by the number of bomb films.
 
 
 
Graph 3. Professor Shõhõji feels that Graph 2 does not provide enough difference between the bomb films and all films to appreciate different trends. Based on his experience, he felt that multiplying bomb films by a factor 30 gives a more useful graphic representation.