If choosing a diet were roulette, your strategy would be to spread your bets across the table, right? It’s true that you’d have a better chance of winning each individual spin than if you bet everything on one number, but in the long run, neither strategy has better odds.
I don’t like to gamble unless I have more information about outcomes. Sure, there are probably undiscovered vitamins, but if we don’t know what they are, our decisions are purely speculative. It’s even possible that a diverse diet could make us deficient in an unknown vitamin, because all choices entail trade-offs. For example, maybe eating a dozen different vegetables makes us eat less corn, and corn just happens to be rich in that vitamin. How could we know? My point is, I don’t think we can make sound decisions based on an absence of information.
I’m actually much more convinced by the fact that your grandmother used to eat greens, and that greens have commonly been eaten in cultures throughout the world. I know many people get by just fine without eating vegetables, but I would feel a little uncomfortable eliminating them entirely, just like I would feel uncomfortable eliminating animal foods or starches.
I guess I’m curious how you define variety, because we probably don’t disagree all that much. To me, adequate variety means eating a representative from each major type of food, but not several instances of each type. If you eat beef, you’re fine without other meats. If you eat rice, you’re fine without other grains. Etc. How much variety do you think is necessary?
Despite our minor disagreements (which seem largely academic), I do agree with your middle ground approach.