The issue isn't that we can't "represent real street results", the issue is that manufacturers took deliberate steps to alter car behaviour with the intention of obtaining better test results than would be present outside the tests.
It's the same with smartphone and notebook batteries. Tell me, when was the last time, you actually got 9 hours of battery live (in case of a MBP)?
Don't get me wrong, i see it the same way as you are, but this is an issue that you can find everywhere not just in car manufacturing.
For example, a lot of power supply companies try all kinds of tricks to game the efficiency test systems.
The article points out that 1/3 of improvement in past 3 years to meet regulation requirements was real improvement and 2/3 was just manipulation.
It doesn't matter that people drive cars differently than they are tested in cars, to illustrate it they highlighted that due to those manipulations that difference now can be as high as 50% which is beyond claims that people drive cars differently than when they are tested in labs.
And, if you got a bad battery that doesn't live up, they'll replace it at no charge.
Battery life declines with age, however.