Yes, and that was more revealing than averaging the sheet. One assembly has an in-fixture alignment tick, an out-of-tolerance released reading, a correction operation, then a final pass. The supports for the released checks are documented too. Thank you, Ravi; I now have a usable sequence, not a defensible correction rate.

Keep that example intact when you present the estimate. It demonstrates real extra work in that case; the incomplete rows should remain incomplete rather than becoming either passes or failures by association.