Using this information often reduces the uncertainty to 15-25 per cent. Nearly any mineral material which has been heated above 500C at a time one wishes to know is a candidate for TL dating. Porcelains, being nearly vitrified, are a special case requiring a fairly large solid core sample, and TL dating of intact objects is not recommended because of the damage caused by sampling.
Most porcelain dating is done for insurance purposes on broken objects.
Some clays are hardly thermoluminescent at all; some may not have a straight-line relationship between dose and TL; spurious luminescence due to chemical or pressure effects may mask the radiation-induced TL; occasionally, a condition called "anomalous fading", where part of the TL is unstable, may lessen the accuracy of the dose measurement.
Generally speaking, when a sample is drilled and there is no information available about the burial environment, one may expect up to 40 per cent uncertainty.
This is adequate for the purposes of authentication where the question is whether the piece was fired in antiquity or recently; it will not differentiate, say, between a classic Greek terra cotta and a Roman copy.