That doesn't result in zero energy efficiency. The energy efficiency is only zero when the heating is insufficient and then it is not the energy that you are putting into he battery that is the problem but the energy needed to keep it warm.

The energy required for charging is the energy input required to fill the battery, plus whatever additional energy is needed to condition the battery. If 50% of the energy you use for charging (which you pay for) goes towards conditioning the battery, then then the energy efficiency of the vehicle is 50% times whatever efficiency you get through the drivetrain. When it gets cold enough, all of the energy draw goes towards conditioning the battery, so the car doesn't charge at all. Since none of the input energy goes towards moving the vehicle, the efficiency is zero, though if you wait long enough it'll take a low value as the temperature increases and the vehicle begins to take charge. It's probably more correct to say that the energy efficiency of the vehicle approaches zero as the temperature drops.

This is different to a standard car, because losses during refueling are negligible, but losses in the drivetrain obviously aren't. And of course for both types of vehicle you have additional upstream effects, e.g. production and transportation of fuel versus production of electricity, production of the vehicle, etc.

The reason I raised the issue is because I hear a lot about well-to-wheels efficiency of BEVs, but I've never heard anyone take into account the effect that conditioning the battery has in certain climates. It's one thing to say that the well-to-wheels efficiency of a BEV is around 60% compared to 16% for ICEVs, but if you're neglecting a large multiplicative efficiency hit in your chain, things like hybrid electric vehicles suddenly compare a lot more favorably.

And if you're comparing mileage on the basis of miles per gallon using energy input at a pump or cable that you paid for, it seems like a reasonable question to ask. Is that mpg figure calculated using the energy of the battery, or the energy that was paid for?

Conditioning the battery isn’t a year round thing 20% more energy for 2 months a year is 1.7% efficiency loss across the full year. It’s also zero in a heated garage.

But again, as soon as you move up the chain on one car you need to move up the chain on the other or you’ll get as distorted picture of the tradeoffs.

I acknowledged upstream effects, but this is specifically about accurately accounting for the energy input required to run the car. If I purchase X units of energy, whether it be in the form of fuel or electricity, how much of that energy is actually going into the battery, and how is it affected by climate? It's one thing to get an answer from California where it doesn't really get cold, quite another if you rely on on-street parking in, say, New York.

It’s still dramatically in favor of EV’s.

NYC doesn’t really get cold enough to matter, average temperatures in January are above freezing. So you’re really talking about a subset of unusually cold days where it’s even slightly an issue.