I always laugh when people use that chart as an example of "misleading graphs"
The general assumption with most of these data presentations is you have some understanding how to read scale. They aren't going to show it from the perspective of 0-100 because it would be unreadable.
If you want a "fair" way to look at it, take a look at this:
Your 96.8% efficiency on the K&N maps to just shy of a 3.4 sigma process...you are talking 32,000 defects per million opportunities.
Compare to the AC Delco at 99.93%, and you have a ~4.7 sigma process, a mere 700 defects per million opportunities.
In this case, your "defect" is a piece of crap that shouldn't be there (ie, dust, dirt), and you might be surprised at the high number of "opportunities" that appear over the course of a drive.
You can debate all day whether or not an 45 TIMES the crap getting past the filter is a big deal or not all day (and I won't bother because it isn't worth my time), but trying to downplay the number as "just 3%" is misleading. "Just 3%" was the difference between a Motorola that almost went out of business in the 80's, and the Motorola that was a major player in the 90's. Percentages are inherently misleading when viewed by most people, especially regarding high-capability processes (like filtering). We could, of course, illustrate with DPMO, but I imagine those results would appear even more skewed.
|