I thought this thread was mainly interested in variations in offset, or, what's your point?

Perhaps it's just that final inspection is easily able to detect and weed out any forks that show that particular defect, or that the process controls are able to effectively prevent that defect. As I said, I am not a composites expert.

In your previous post you say that the same process controls should control accurate offset as well as accurate OLD at the fork. It seems intuitively correct, but I do think the tests for the two are not equal in labor or production line time, nor are the product requirements the same. Every fork released to retail sale needs to readily and correctly accept the QR front wheel. We seem here to have agreed that a few mm of variation in offset is not very important to the product, so correct lateral spacing would be more important than correct offset. Hypothetically, if it takes more effort to check offset of each fork than to check lateral dropout spacing (a hand-held go/no-go gauge which could itself be a very cheap injection-molded part, I presume) of each, then perhaps only every hundredth fork will be checked for offset, where all will be checked for lateral. So effectively zero forks will be released with a lateral spacing error, and some forks with an offset error could be deemed acceptable, even if they are not at the nominal value.