ABSTRACT We show that the radial velocity variable star in the black hole (BH) candidate NGC 1850 BH1 cannot be a normal ${≈}5\hbox{-}{\rm M}_{⊙ }$ subgiant, as was proposed, but is an overluminous stripped-envelope star with mass ≈1 M⊙. The result follows directly from the star’s observed radius and the orbital period–density relation for Roche lobe-filling stars: The star’s density, as constrained by the observed ellipsoidal variability, is too low for its mass to exceed ${≈}1.5\, {\rm M}_{⊙ }$. This lower mass significantly reduces the implied mass of the unseen companion and qualitative interpretation of the system, such that a normal main-sequence companion with mass $2.5\!-\!5\, {\rm M}_{⊙ }$ is fully consistent with the data. We explore evolutionary scenarios that could produce the binary using mesa and find that its properties can be matched by models in which an ${≈}5\hbox{-}{\rm M}_{⊙ }$ primary loses most of its envelope to a companion and is observed in a bloated state before contracting to become a core helium burning sdOB star. This is similar to the scenario proposed to explain the binaries LB-1 and HR 6819. Though it likely does not contain a BH, NGC 1850 BH1 provides an interesting test case for binary evolution models, particularly given its membership in a cluster of known age.