When I write something that contains formulas, I only number the ones that I refer to myself. However, I found that some books or papers just number ever single formula. And that makes it really easy to cite individual formulas.
Is there some general rule to this?
The guideline for this is to use whatever method the journal to which you are submitting asks you to use.
As you indicate, many journals do this automatically, for exactly the reasons you suggest—to make it easier for others to cite and follow along your work. However, not all journals do this, and as a result, it's better to follow their guidelines instead of improvising.
