Are they still more efficient if you factor in the whole chain of fuel->wheel rotation? Converting electricity to battery charge, to movement surely shaves off at least 15%.

They are. Remember that you also have to factor in the whole chain of crude oil underground -> wheel rotation if you want a proper comparison. Transporting and refining gasoline isn't free either.

Car engines have pretty bad efficiency. You're looking at something like 25-30% typically, and that's peak efficiency. Car engines usually run well off peak efficiency because they're not at optimal RPM or power output, and then you lose even more when braking. Add on another substantial portion for refining and transport.

A modern power plant can be 60% efficient or more. You lose about 6% in transmission, and then another 5-10% in the rectifier, battery, inverter, and motor.

If you factor in the chain from crude, you also have to factor in the mining and transportation of coal. I don't think that's usually done when calculating the CO2/kWh.

It needs to be if you want to make a meaningful comparison. If you search for the lifecycle emissions of coal you'll find this accounted for. That search will also give you non-zero values for things like nuclear and wind, which is also useful. (Although a bit hard to say for sure, since in a hypothetical world powered exclusively by nuclear or wind, the values would be lower.)

> They are. Remember that you also have to factor in the whole chain of crude oil underground -> wheel rotation if you want a proper comparison. Transporting and refining gasoline isn't free either.

Do you? I mean... we're still going to refine petroleum if we go to electric cars completely, if only for industrial purposes. Even then I think it's going to be a while before anybody proposes electric passenger jets or electric cargo freighters, so there's still going to be refining and oil exploration. I am not really sure it is correct to assign 100% of the cost of that to cars.

Who said anything about assigning 100% of the cost to cars? You break it down into a per-gallon figure, and use that to assign the cost to cars.

It's the same thing with electrics. You don't assign the entire energy lost in the power grid's wiring to EVs. You assign them the percentage loss on the energy they consume.

To obtain a gallon of gasoline costs a certain amount of energy which implies a certain amount of emissions attributable to that gallon. If you're looking at the entire efficiency of a gasoline car in terms of miles per unit of CO2 emitted, you need to account for that. The same goes for a kilowatt-hour of electricity and accounting for the miles-per-CO2 of an EV.