[mindslight]

I just think this whole topic gets derailed in many ways by the lure of comparing things to resistive heating (eg the original article), and using parity with resistive heating as a reference point for practical heat pump source temperatures is an instance of this. The benefit from a heat pump has left long before COP = 1. I'd say a COP of 3ish (modern modcon boiler efficiency divided by common gas power plant efficiency) is a straightforward honest appraisal of where a heat pump starts to be less than ideal.

Of course I recognize low temperatures are merely one small portion of the time. Residential heat pump system design has generally taken advantage of this by having backup resistive coils. It would be much more efficient to fall back to a gas furnace, but for the capital cost. So an actual comparison necessarily takes into account expected temperatures over the day and the heating season, likely nighttime setback, thermal mass, etc. Which certainly isn't easy to button up into a simple comparison, but the point is resistive heating is a red herring!