Fable
Two test pieces in the porcelain kiln: one hand-marked by the master, one randomly inspected by the apprentice
Before porcelain kilns ship, there are two inspections. The first is batch sampling done by the apprentice team—each kiln produces two hundred pieces, and the apprentices sample one out of every ten by kiln position number, checking for bubbles, cracks, and uneven glaze. The sampling results are written in the daily kiln report: this kiln's pass rate is ninety-three percent. The second is the master's calibration inspection—on the day of firing, twenty pieces are drawn from the same batch, and each one is examined spot by spot under a magnifying glass, compared against the color chart for glaze color, and measured at the foot with a vernier caliper. The calibration report is thicker and slower, averaging one shichen per report.
Once, on a batch of official kiln orders, the apprentice's sampling gave a 95% pass rate, and the owner shipped the goods to the official kiln's acceptance camp based on that number. The acceptance officer inspected according to court standards—stricter than the old master's calibration standards—and found seventeen hidden hairline glaze cracks, and the entire batch was impounded. The owner went back to confront the apprentice, who said, I sampled by kiln position, and everything I sampled did pass. Where was the problem? The kiln positions were unevenly distributed: the three rows nearest the furnace door already had a higher defect rate, and the apprentice's sampling happened to skip right over them.
The master's calibration inspection did the same—he did not sample by kiln position number, but took two pieces from every row. In the three rows nearest the kiln door, four of five pieces had hidden hairline cracks. The calibration report showed a pass rate of seventy-two percent. The apprentice's kiln reported ninety-five percent, the master's calibration seventy-two percent, a gap of twenty-three points. That missing twenty-three percent was exactly the area where kiln temperature was uneven.
The owner changed two things. First, the master's calibration inspection became a gate that had to be passed before shipment—no matter what the apprentice's sampling report said, if calibration did not meet the standard, goods did not ship. Second, the calibration inspection results were archived separately and not combined with the kiln report in the same calculation. The two reports each had their own use: the kiln report looked at daily production fluctuations, and calibration looked at the critical quality red line.
Later, an out-of-town merchant came to discuss bulk supply, and the owner showed him two reports. The merchant asked why the numbers in the two reports differed so much. The owner explained: the kiln report is a coarse sieve; ninety-five percent means "looking across the board, most are fine." The calibration report is a precision magnifying glass; seventy-two percent means "judged by the strictest standard, three out of ten have hidden risks." The merchant asked which one to sign the contract by. The owner said that if the contract says "pass rate," then it must specify which report's standard is being used—the same kiln can report ninety-five percent under sampling inspection and seventy-two percent under calibration, a huge difference in numbers.
After the merchant left, the owner posted a notice at the kiln: spot checks govern the production rhythm — a daily quick report makes it easy to adjust the fire, swap the blanks, and arrange kiln positions. Calibration governs the shipping promise — every kiln must be inspected, to the strictest standard. One is a frontline tool, the other a customer commitment. Sending spot-check numbers out inside a calibration report shell follows not the firing line but the relationship line.