The link in the first post by OP.
That link doesn't go to the full story anymore.. And yesterday the only detail in there was that the details would be released today (uhm)
this story has details:
In both testing modes, Mazda did not find any sort of *improper alteration or falsification of test data.*
However, the company did detect speed trace errors, which refer to situations in which the car being tested deviates more than the permitted amount from the pre-established speed trace pattern as described by the test mode. These speed trace errors were discovered in 72 cases out of a total of 1,472 cars that were tested under the aforementioned JC08 cycle.
Following a more in-depth investigation, Mazda detected two causes for the mishap. The first one is about the system not being correctly programmed to automatically invalidate results containing a speed trace error. In addition, these test procedures were incorrectly organized as detecting the speed trace deviations became the responsibility of each individual inspector.
Mazda has already re-examined all the test data and the results are: *No effect on specification fuel economy and emission figures.* The company goes on to say it hasn*t encountered any issues during the WLTC test cycle.
Put down the pitchforks everyone. A minor programming Error in a test machine and a small procedural error is not considered "cheating".
Looks like the only real cheating going on here is the made up news headline in the original post from yesterday.