Slow down there. You're comparing the complete-cycle efficiency for petroleum to just the end-stage efficiency for electric.
You seem to have not noticed what this article is about. It's about making fuel from electricity and then giving it to cars. Both sides start with the same feedstock: electricity. So it doesn't matter how efficient the electricity was to make because it affects both paths equally.
But let's switch back to your "scenario that I want to talk about that's not the one in the article"
Slow down there. You're comparing the complete-cycle efficiency for petroleum to just the end-stage efficiency for electric. That electricity needs to be made somehow. Toss in 40% efficiency for coal plants (we'll leave out pumping/mining and fuel transport costs for now, assuming they're similar for oil and coal), battery charging efficiency of about 75% [futurepundit.com] (discharge efficiency is unspecified, but since the EPA mileage estimates are based on battery capacity it's safe to ignore it), and the 85% motor efficiency you've specified, and suddenly your EV is .4*.75*.85 = 25.5% efficient. Same as a diesel.
I don't see that figure in your link, and I don't really need your link because I'm familiar with the numbers already. It depends on what you mean by "charging efficiency". The US grid averages about 8% distribution losses, plant to breaker. Li-ions are over 99% efficient at slow charging, but depending on the type can drop a few percent in faster charging scenarios, and in an extreme situation down to the lower 90%s. The charger itself has some losses, if I recall correctly from the breaker they're usually 92-94% efficient. So a good middle of the road number is more like 84%.
Also note that EVs automatically also function as hybrids: they regen and don't "idle".
Their EV is cheaper to operate not because the EV is more energy-efficient, but because coal is so much cheaper than gasoline
Coal is of course the dirtiest widely used power source, and its usage is declining in most first-world countries. Natural gas and wind have the highest growth rates. The most efficient combined cycle natural gas plants are upwards of 60% efficient, although that's not an "average" efficiency, but even old plants are generally over 40%. Efficiencies on things like wind, solar, etc are of course not particularly meaningful, since you're not burning a fuel. Nuclear has a low efficiency, but again, that's not particularly meaningful.
Even putting solar panels on your roof and amortizing the costs in most climates makes running an EV cheaper than gasoline. It's not because coal is somehow ridiculously cheap. It's because oil is a really expensive energy source per joule.
Wind is only about twice the costs of coal
If this was true, people would be churning out new coal plants, not wind farms.