I think that's exactly what
I think that's exactly what CR was trying to point out, that your actual mileage may vary. In their testing, in which they allegedly duplicate real world driving more accurately than the bizarre cycles the EPA uses, they found that the car didn't perform up to EPA-rated expectations. Especially with hybrids, there are infinite control strategy tweaks that can change the way the car responds to different driving conditions. Sometimes automakers, unintentionally or not, create a control strategy that responds to the specific driving cycles better than actual driving. Ford's hybrids had a huge problem with this recently. Not malicious on CR's part, just an important objective observation.