The Mediterranean pattern is the most consistently recommended diet in research, and also the one most often reduced to a shopping list. The gap between those two things explains a lot.
It was observed, not designed
The pattern was documented by researchers comparing populations, and it described how certain regions ate in the middle of the last century rather than a plan anyone had devised.
That origin matters. A designed diet has a mechanism in mind, while an observed one captures whatever happened to be true of a place, including things nobody thought to measure.
What was recorded included the food, but the food arrived inside a way of living that the records could not fully separate from it.
The food list is only part of it
The commonly cited components are vegetables, legumes, whole grains, fish, olive oil as the main fat, and modest quantities of meat and dairy.
Less often carried across are the structural features: meals eaten with other people, cooking from raw ingredients, seasonal availability and long gaps between eating occasions.
Physical activity in the original populations was also built into daily life through work and walking rather than added as exercise, which changes the energy balance the diet sat within.
Ingredients change when they travel
Local produce eaten in season differs from the same items shipped and stored, and the difference shows in fibre, in variety and in how much is eaten fresh.
Olive oil bought as a supermarket commodity is not consumed in the quantities or the roles it had where it was pressed locally and used as the default cooking fat.
A pattern reconstructed from imported components in a different food system produces something recognisably related but not identical.
Adherence scores measure a composite
Studies rate participants on a scale counting how many pattern components they meet, which produces a single number covering many separate behaviours.
A high score can be reached by different combinations, so the score does not identify which component is doing the work, and it may be that no single one is.
The interacting-whole explanation is plausible but hard to test, because trials that isolate one component lose the interaction being investigated.
Why it is still the default recommendation
Despite the difficulty of transplanting it, the pattern retains the broadest evidence base of any dietary approach and it is unusually easy to sustain.
It excludes almost nothing outright, which removes the failure mode that ends most restrictive plans, and it maps onto ordinary meals rather than replacing them.
Partial adoption also appears to carry partial benefit, so the pattern degrades gracefully in a way that all-or-nothing regimes do not.