[joshlemer]

But it is evidence for causation.

[lucketone]

Is it?

[joshlemer]

Yes. If A is correlated with B, is A more likely to cause B than if A is not correlated with B? Clearly yes.