We are talking about mass production items, not hand-made showcases here. I rarely encounter a defective sample right out of the box, but it *does* happen. it is quite hard to make an estimate of the likelihood of getting a problem lens, because the input data is highly skewed. People make a lot of fuss about problem lenses, whilst happy owners don't contribute. The ratio of these groups may be 50:1 or 100:1 or even higher in favour of good products, but we cannot give valid estimates. My gut feeling is that we are looking at 100:1 or better, but that is only based on my own samples received for tests and reviews. The last lens I encountered with obvious decentering was a 70-200 f/4 Nikkor and the decentering was quite mild and only causing some trouble for distant scenes at the long end of its focal range, so might have passed unnoticed by many users. From this I cannot at all conclude the 70-200/4 is a flawed lens, only that my single sample had an issue.
