12-17 days after the power failed, the contractor measured temperatures in houses that had had neither power nor occupants during that time. "None of the air or surface temperatures in the 31 houses inspected had fallen below 0°C."
There is a proviso: "In very cold temperatures (for example, -30°C), high winds or houses without basements, the rate of cooling will be much faster than observed in the southern Quebec houses of this study."
Now, I lived for a year in a house which was poorly insulated, and where you could feel the draft through the window frame. When it was -15F outside 2L pop bottles would freeze on the floor of the pantry even when the heat was on, and when the furnace went out during Thanksgiving weekend the inside temperature quickly dropped. Even then, it didn't match the freezing outside temperature over a couple of hours. I can well believe that your experience is true, but it is atypical.
On the other side of things, Passivhaus designs (see http://en.wikipedia.org/wiki/Passive_house ) give an example of what can be done with good design, for houses designed for freezing weather. They need only 10W/m² of heat, which can often be met by normal activity (people, computers, screens, coffee makers, etc.)
It's possible to have south facing windows in Germany which let in more solar energy in winter than is released by heat loss.
Add in some passive solar and you can keep a house in the continual US warm without active heat sources fairly easily. The major issue is it stops looking like a traditional house and costs more up front, but it’s not a major issue.
PS: ~6 hours * 1kw/m^2. Take a 30' x 30' roof and that's ~600 kwh per day or an average of 24kw of heating. Though you do need to keep the now off the roof in the winter, but a steep enough rough does that fairly easily.
Ideally, if you can build your entire house below the frost line, you're laughing - but this isn't always feasible.
My understanding is if you live near a stream or pond it's a great way to make heat pumps far more efficient, but otherwise just having a large rock slab in the middle of your house combined with solar heating is far cheaper for heating, but not that useful for cooling if your area does not cool off at night.
Of course, all this is reliant on knowledge of the local geography and climate before you purchase and get set up. If you're limited to where you are right now, you're kind of stuck with what you can put on your property right now. Which could be as limited as running on-grid but growing your own food.
In my experience, a house heated to 80F can be 30F within a matter of a couple of hours, depending on the number of windows, temperature outside, and insulation levels.
I certainly don't think that's a solution to winter problems in cold-weather climates.
As for storing the heat, I'm not really sure what you mean. Surely it would be inefficient to generate extra heat during the day and store it, as opposed to just creating some in the evening?