The Final-over-Final Condition (FOFC) (Sheehan, Biberauer, Roberts, and Holmberg 2017) purports to be a universal word order constraint. In this article, we challenge this claim and demonstrate that the FOFC is of a statistical nature: It is a relativized¸ non-absolute version of cross-categorial harmony, where only a head-final projection dominating a head-initial projection is ruled out among the disharmonic configurations displaying different head directionalities. The cross-categorial generalizations in the World Atlas of Languages, referred to by Sheehan et al. (2017) in order to illustrate the crosslinguistic validity of the FOFC, are shown to be useless for determining head-directionality, given that they systematically gloss over functional categories. The Mandarin Chinese head-final split CP dominating a uniformly head-initial extended verbal projection and TP serves as a case study here, because it has challenged the FOFC since its very beginning. The numerous efforts to make the Chinese CP “FOFC-compliant”, the latest being Biberauer (2017), are shown to be unsuccessful and to ignore well-established principles of Chinese syntax. The data from Chinese thus add to the evidence from other languages likewise undermining the FOFC.