Originally Posted By: Garak
I would agree with that comment about statistics. In fact, if the error bar is bigger than the data point's magnitude, there really isn't a data point at all.
Yup, Shell said in the Jobbers World article six years ago that is linked above that the standard deviation, which is the standard error of the sequence IVA measurement, is 12.5 µm. Therefore, any sequence IVA result less than 12.5 µm is indistinguishable from zero (0 µm) or no wear at all due to statistical error. Therefore, any oil that scores 12.5 µm or less can safely claim that no other oil performs better than theirs in sequence IVA. They cannot claim the reverse though, which is that our oil performs better than all other oils, which would be wrong. For example, Castrol Edge is shown to have scored 10 µm in the article and they can also claim that no other oil performs better than theirs.
In fact, in the article, Shell went very far and said that results within three standard errors, which is 37.5 µm, are indistinguishable. I think this is overkill though, as this is not like looking for a new particle in a particle accelerator.
We don't know whether Pennzoil's current result is 12.5 µm or less or 37.5 µm, but their claim is a marketing gimmick that relies on either the one-standard-error or three-standard-error argument, and there are many other conventional and synthetic oils which score similar in sequence IVA, and they can all claim the same.