The rule as usually stated
It appears everywhere, often as a table: green tea at seventy or eighty degrees, oolong at ninety, black tea at boiling, with the implication that the number follows from the category, and sometimes with a percentage of oxidation attached to make it look derived. It has obvious appeal. It is memorable, it maps onto the six categories, and it is right often enough to feel reliable. It is also a correlation that has been promoted to a mechanism, and the promotion is where the trouble starts.
Why it seems to work
The correlation is real. Unoxidised teas are, on average, made from younger and more tender leaf, are less thoroughly worked, and carry more of the compounds that turn harsh with heat. Heavily oxidised teas are, on average, more robust and better able to take boiling water. So a rule that maps category to temperature will land in a reasonable place most of the time. The problem is that the underlying causes are leaf structure, tenderness, roast and processing — and oxidation happens to travel alongside them rather than to cause the temperature requirement.
The counter-examples are not obscure
Gyokuro and an ordinary sencha are both unoxidised green teas and want water that differs by twenty degrees or more. Silver needle white tea is barely oxidised and takes water far hotter than any green tea of comparable delicacy. A lightly oxidised high-mountain oolong wants near-boiling water despite sitting at the green end of its category, while a green tea sitting at the same nominal oxidation wants nothing of the kind. A heavily roasted oolong and an unroasted one of identical oxidation want quite different treatment because the roast changed the leaf. These are not edge cases; they are among the best-known teas in the catalogue.
The percentage problem specifically
Oxidation percentages — thirty per cent, seventy per cent — are quoted constantly and are not measurements of anything. There is no standard assay that produces them and no agreed definition of what would be measured, so the figures circulating in tea writing are conventions and estimates repeated until they acquired the appearance of data. TeaHQ does not record them anywhere; the catalogue uses a descriptive vocabulary instead. A temperature rule built on top of a fabricated number inherits its confidence without inheriting any support.
What actually sets the temperature
Several things, and they need to be assessed together. How tender the leaf is — buds and young leaves want cooler water than mature leaf. Whether it was shaded, which changes the composition in a direction that rewards low temperature. Whether it was roasted, which needs heat to open. Whether it is compressed, which needs heat and time to penetrate. How small the particles are, which changes the rate at any temperature. And how the tea was fired or steamed, which affects how readily it gives up what it has. Oxidation is one input among these rather than the master variable.
A better heuristic
Ask what the leaf looks like rather than which category it belongs to. Tender, downy, small and pale: go cooler. Large, dark, twisted, roasted or compressed: go hotter. Shaded Japanese green tea: cooler than you think. Anything you would describe as robust: boiling. This gets you closer than the oxidation rule does, and it has the advantage of working on a tea whose category you do not know, which is exactly the situation where a heuristic is needed at all.
When you know nothing about a tea
Start cooler and shorter than you think, taste, and adjust upward. That order matters because the two errors are not symmetrical: an under-extracted cup tells you clearly what it needs and the leaf survives to be brewed again, whereas an over-extracted one has already released the harsh fraction and cannot be walked back. Consult the tea’s own recorded parameters where they exist, since a set authored for that tea and that method beats every general rule including the ones in this guide.