Sorry - you quoted me quoting you saying "So if a circuit is added and tests to BS7671 are not done on that circuit, essentially the question is - so what? (So long as a faulty circuit is not extended into a controlled zone.)" so I took "second part of my original question" to mean the second part of that.Difficult to know who/what you are replying to. What you say might be relevant to something I said later on but not to the "second part of my original question" which was
If there is a problem not found with simple tools (lets imagine a loop impedance too high for a short to be detected by breakers), how often does it actually happen? How many RCDs come with parametric failures straight from the box (as distinct from obviously broken)? etc
My Bad.