ASHRAE—an HVAC organization—has data on the coldest and hottest days for areas so that you can design things for the coldest or hottest 1% of the year (4 hottest/coldest days):
I think that if you have an older, leaky/ier, less-insulated house you may need to 'brute force' heating your (probably older) domicile. But if you have a <4 ACH@50 air tightness, and reasonable insulation levels, a good portion of the US population could make do with a heat pump.
Mitsubishi publishes data were they have 100% heating capacity at -15C, which some models being 100% at -20C and -23C:
It is warmer than -16C/3F at Chicago (O'Hare) for 99% of the time (i.e., except for 4 days a year), and warmer than -18.7C/-2F for 99.6% of the time (2 days).
ASHRAE are the folks that publish the heating/cooling standards that are used in building codes for estimate heating/cooling equipment capacities (Manual J) and selecting the right equipment (Manual S).
Here's a PDF with a lot of locations in the US and CA (and other countries further down), and if you look under the "Heating DB" column, you'll find very few US locations that have -30F under the 99% (or even 99.6%) sub-columns:
So unless you're in AK, MN, or ND, long runs of temperatures colder than -20F/-30C don't happen too often. Of course if you have a leaky house with little insulation, you're throwing money out the window/door, so the first consideration for a good ROI is better air sealing and insulation.
I think the comment was saying below 30F and below 10F. Much warmer than you're saying.
Also..
> It is warmer than -16C/3F at Chicago (O'Hare) for 99% of the time (i.e., except for 4 days a year), and warmer than -18.7C/-2F for 99.6% of the time (2 days).
If my heat doesn't work for those days, I'm kind of boned. Four days per year without a working heat pump? That's a mess.
At face value, then in the worst case that's just 4 days per year of using resistive heat to keep a home warm.
Which is, of course, very expensive to use -- but it's only expensive for those 4 days. Resistive heat can be avoided for the other 361.2425 days in a year.
In the US (as of August of 2025), the average price of residential electricity per delivered kWh is $0.1762 [1].
If using resistive heat averages 4kW during each of those 4 days (it's probably either more than that, or less than that, but ballparks are ballparks), then that's about $16.92 for each of those days. Or: $67.66, per year.
> At face value, then in the worst case that's just 4 days per year of using resistive heat to keep a home warm.
The design philosophy for using 1% is that you may end up having to run your heating (or cooling) 24/7 to keep up with temperature delta between outside and desired inside, but it will keep up with the demand.
The rest of the time (99%) the mechanicals only run intermittently. Also note that the 1% would not necessarily occur every year: it is just the historical average. The charts also have the 0.4% extremes if you want to be extra conservative, but most building codes specify 1% because that is what experience has shown is a good trade-off.
Part of the process (in the US) is to use what is called the Manual J to determine/estimate/calculate how much energy is needed to maintain a particular temperature (typically ≥70F/21C in winter, ≤75F/24C in summer):
> The Cooling Design Day is effectively the "worst case" day for your air conditioning loads. The "worst case" hour of this day determines equipment capacity, fan sizes, and subsequently duct sizes. This largely impacts first cost. The Design Hour also impacts peak KW demand which often has a huge impact on the utility bill.
So gas hash higher reliability and is cheap for the times you need heat the most, whereas heat pumps might not work and are not cheap at the times you need them the most?
I’ve had a gas furnace keep me and the water heated multiple times in a cold weather power outage.
> So gas hash higher reliability and is cheap for the times you need heat the most, whereas heat pumps might not work and are not cheap at the times you need them the most?
The major manufacturers have systems that will use the heat pump when the temperatures are not 'crazy', and kick in fossil at a certain point:
Depending on the cost of power and fossil fuels, you can program it to switch over once the COP becomes too low to justify running up kWh on your meter.
But whereas in the past heat pumps would have their COP drop around 40F/5C, modern systems can be fairly efficient at much lower temperature nowadays:
That's one of the older style units. Starting in 2007 when Mitsubishi introduced their "Hyper-Heating Inverter" heat pumps, and continuing with Fujitsu and Daikin following with similar technology in the 2010-12 timeframe, and others a few years later, heat pumps got way better in the cold.
Mitsubishi's maintain 200%+ efficiency down to -4℉ (-20℃) and 150% down to -22℉ (-30℃) [1]. Only a few towns in the continental US get below that, and even those aren't going to get cold enough long enough to make it worth it an an all electric home to switch to your emergency electrical resistance heating.
Their capacity doesn't start dropping until you get down to 23℉ (-5℃), dropping to 76% at -13℉ (-25℃).
Though it’s worth noting that that first 2 ton rated unit is putting out 0.5 tons (6k BTU/hr) at that temp and rating.
That’s not going to be particularly helpful for a structure that needed 24k BTU/hr during warmer temps, meaning the owner of the unit is likely mixing in a lot of 1.0 BTUs to meet the heat loss at -13°F.
> Though it’s worth noting that that first 2 ton rated unit is putting out 0.5 tons (6k BTU/hr) at that temp and rating.
I just did a quick search for "all" units and sorted the result list/table by COP@5F. If one was actually shopping/designing a solution then a more nuanced search criteria would be used.
Further, you'd probably want to do a (US ACCA) Manual J calculation to first determine how much energy is needed (j = joules)
I've got one about 8 years old, and it does just fine down to 0°F (it hasn't gotten colder than that here). It doesn't even have any kind of auxiliary heat.
It's fine. The only difference when it's super-cold is that the air coming out of it isn't as warm, so the heating cycle stays on for a longer proportion of the time. But it keeps it 70°F inside no problem at all.
Mine struggles if it gets below 30, and might as well not exist below 10. They're not great at low temps.