Your first paragraph presents a straw man. Sure, in extreme cases, reproducing a bug is not mandatory. Again, the author does not claim that reproducing a bug is mandatory, but rather that it is a useful practice.
It is not a virtue to try to base your work on the least amount of information.
You say that "when the hard problems come you will be ready". But really when the hard problems come you will look at them assuming that the logged information is enough to solve them. No programmer can know ahead of time what information to log, so by purposefully blindfolding yourself from experiencing the bug directly, you might miss the bigger picture.
Indeed, I can't think of a reasonable logging mechanism that the author could have thought of ahead of time that would have helped with this particular bug. Emphasis on "reasonable".