I'm working on some problems that demonstrate some simple implications. The logic seems to be very different from the way I'm used to using it in everyday language. I'm not sure what assumptions I am allowed to make to show that the statement is true or false. For example, can I consider hypothetical worlds in which a statement I know to be false is actually true? Below is an implication that I need to show can or cannot always be true.
Every good boy does fine $\Rightarrow$ Some bad boy doesn't do fine.
- True, assuming every good boy doesn't do fine.
- It could be false though, assuming every good boy does fine is true and some bad boy doesn't do fine is false.
If I'm allowed to assume anything I want how can an implication always be true? What am I missing?

$A \Rightarrow B$ doesn't translate well into English ever. In natural language you assume motivation behind statements, so a casual statement $A \Rightarrow B$ often has a suggested "and $A$ is true". It is also very hard to avoid hangups with quantification (ambiguously assuming the natural language speaker meant some kind of "for all A"). Implication in natural language is nasty.
For a statement $A \Rightarrow B$ I suggest translating it as $(\text{not } A) \text{ or } B$. It will be much easier for you to think through. So
becomes
So if you assume "every good boy doesn't do fine", then the statement is true, since the first half of the "or" is true.
If you assume every good boy does fine and every bad boy doesn't do fine then it is also true, since the second half of the or "some bad boy doesn't do fine" is satisfied.
If would be false in the case of both halves of the "or" being false:
Aside,
The opposite of "every good boy does fine" isn't "every good boy doesn't do fine". It is "there is a good boy who doesn't do fine". You only need 1 boy to be a counter example to the "every".