Koberlein's explanation is okay but I read Burns's "Keep velocity constant on z-axis" as an instruction rather than an assumption. In that light the error doesn't appear so subtle (it manifests as a simple error of "conservation arithmetic").