It's interesting. Given the context (Krauss' larger argument) the point is that among the millions of pieces of scientific data which the human race has carefully collected, prudently analyzed, ruthlessly abstracted, reduced, synthesized with other data, etc., we have never encountered anything remotely resembling a valid miracle that has withstood any degree of scrutiny. Part of the problem may be the definition of the word. Just like "purpose" is meaningless (and even cognitively harmful) when applied outside the context of human affairs, the word "miracle" doesn't have much meaning outside of a theological context. Absent that, in a secular context 'miracle' tends to simply indicate 'against all odds' or 'nearly impossible.' Then again, on the other hand, many theologians have declared faith through miracle to be inferior to a faith arrived at by some other means which they prefer, whether through revelation or election or reason, etc.