Uh, ok...
That's the so-called scientific method -- which is a terrible name, not because it doesn't work but because it encourages the mistaken belief that anything that does not use it is not a science. The intro to the wikipedia article says the following:
Science (from Latin scientia, meaning "knowledge") is a systematic enterprise that builds and organizes knowledge in the form of testable explanations and predictions about the universe.[1] In an older and closely related meaning (found, for example, in Aristotle), "science" refers to the body of reliable knowledge itself, of the type that can be logically and rationally explained (see History and philosophy below).This is indicative of either (a) what linguistics would call drift, or (b) the same systematic redefinition of words to control thoughts that the liberal/progressive/asshats employ: "winning the battlefield by controlling the language."
There is no real reason to discard the older definition [underlined] should be discarded in favor of the new [italicized]; and, indeed, a lot of damage can be done by trying to do so -- if mathematics is not a science then how can science use its branches (logic) or techniques (deduction, induction, etc)... indeed how can something be falsifiable [by science] when you reject from 'science' the mechanisms of reasoning?
Also, as a counter-example of math not having predictability and (and even falsifiability) I would submit the very computer you are using -- for the very software it runs bridges the gap of mathematics's specifications to the program which is the implementation. -- I am a Computer Scientist, and I can reason about a program proving and disproving attributes and situations [debugging].
Uh, ok...