Summary: For chemical analyses with inaccurately reported H₂O(+) or F, calculation of an amphibole formula on the basis of 23(O) after discarding the reported H₂O(+) is in most cases as unsatisfactory as a standard calculation including H₂O(+) based on 24(O, OH, F). The sum of the cations in X, Y, and Z groups most closely approaches theoretical values in a 23(O) calculation, but only by virtue of compensating and compounded errors of opposite sign. Realization of the ideal formula by either method without additional data is not an infallible criterion of accuracy of the analysis or the correct site occupation. These conclusions stem from the observation that although H₂O(+) may be incorrect, most analyses add up to 100.00 or greater; thus compensating errors in metallic oxides are implicit.