ModuleSet.calculateEntropy()

[small-pineapple]

It's return is "-1f / total * entropySum + Mathf.Log(total);"
What does this mean? Normalization?
The final effect will be better if just use "-1f * entropySum" or "entropySum", the result generated by normalized version is too flat in height

[marian42]

This formula for entropy is a bit counter intuitive, but it can be derived like this:

That being said, you don't have to use entropy, if you get better looking results with a different formula, that's great!

[small-pineapple]

So ModulePrototype.Probability is relative probability or weight.
However, the entropy value is dominated by the size of p (ModuleSet.Count), those entropies should have been normalized before compare.