It's entirely possible that the best of all diets, whatever that is, actually does include some supplements. You might have promised not to be a jerk, but I never promised to not be pedantic.

Backtracking a bit--okay. I could see a study of one as being definitive, my favourite example is the first diabetic kid to put on healthy fat pads when insulin for type I was first administered. And certainly as you say, if a claim is that any human on an all meat diet is going to develop scurvy, that doesn't take many subjects to disprove. But while you did say scurvy, you said, especially scurvy. There are a lot of nutrients besides vitamin c, and a lot of nutrients we can be deficient in for quite a while without obvious ill effect. Or that we can be deficient in even before a trial starts. If deficient zinc was lowering my testosterone levels, for instance, and I went from one zinc deficient diet to another, that might just continue the status quo, whereas if I started on a zinc-sufficient diet and switched to the deficient one, over time I might develop symptoms.

The way nutritional requirements are defined--take a sufficiently large group, see how much of a nutrient it takes to keep most, close to a hundred percent, from being in a negative balance in the case of something like protein or minerals like calcium or magnesium, or in the case of various vitamins to avoid symptoms of deficiency. Sometimes there's quite a wide range of intakes that look to be adequate on an individual basis.