And the different measurement procedures and conditions are exactly how the "discrepancy" was found. It seems there was, up to that point, no interest in doing

And the different measurement procedures and conditions are exactly how the "discrepancy" was found. It seems there was, up to that point, no interest in doing the tests "under the real driving conditions." And then there is a "shock" that the engine during the test runs optimized for the given test.