first, we all know there's a reasnonable correlation between epa estimates and our observed results. It's quite good. Second there's been lots of tests on other standards to show they perfrom good real world estimates. The author is nit picking that a specific set of linear combinations of tests hasn't been correlated. That is if you have 5 test that are correlated with real world measurements and you average them it is true that the average has not been tested but logically we can estimate it's error from the other tests.
who writes this crap