Well starving could be considered just a 100% diet rate. Clearly that would be catastrophic so any claim that dieting is beneficial is obviously fallacious and in fact we would be better off eating more.

Looking at behavior in the extreme case is only valid if behavior scales smoothly up to that limit.