The obvious prediction, and why it fails
Everything in brewing advice says that leaf left in water keeps extracting, so an unstrained glass drunk over an afternoon ought to be undrinkable by the middle of it. Enormous numbers of people in China drink exactly that every day, from a glass or a flask, without any difficulty and without any sense that they are getting away with something. The resolution is not that the standard advice is wrong. It is that the advice was written for a western leaf dose and stops applying well below it. Grandpa brewing is not the same method with the strainer left out and the consequences accepted; it is a different arrangement of the same variables, in which the low dose is the load-bearing element rather than a compromise on strength. Reading it as an economy measure — less leaf because leaf is expensive — gets the causation backwards, and leads people to try it with a normal dose and conclude it does not work.
Extraction slows as the water fills up
Dissolution is not indefinite and does not proceed at a constant rate. As the concentration of dissolved material in the water rises, the rate at which more comes out of the leaf falls, and it falls towards zero as the water approaches what it will hold from that material at that temperature. In a high-dose brew, that limit sits far above the point at which the drink becomes unpleasant, so the slowing never rescues you — the tea passes through drinkable and keeps going, and removing the leaf is genuinely essential. In a low-dose brew, the leaf simply does not contain enough to push the concentration past palatable: the glass approaches a plateau and then sits there more or less indefinitely. That is the entire mechanism, and it explains why dose rather than time is the parameter that decides whether a method can work without a strainer. Everything else about grandpa style follows from it.
Topping up is what keeps it moving
The plateau would eventually become dull on its own, since a glass sitting at equilibrium is not gaining anything. What keeps a glass going for hours is that it is drunk down to roughly a third and then refilled, which dilutes the liquor, drops the concentration, and restarts extraction from leaf that still has a great deal left in it. Each refill is a little weaker than the last as the leaf gives out, so the drink declines slowly across an afternoon rather than intensifying — precisely the opposite of what the intuition predicts, and the reason the method is sustainable rather than merely tolerable. Letting the glass run dry breaks the cycle in a way that is immediately noticeable: the remaining liquid concentrates sharply as it evaporates and warms, the leaf sits exposed, and the next fill starts from a much stronger and harsher base. Topping up before the glass is empty is the one piece of technique the method has.
The temperature drift helps rather than hurting
A glass of tea cools continuously from the moment it is poured, and cooler water extracts more slowly and takes a different balance — noticeably less of the astringent fraction in particular. So the same downward drift that quietly ruins a long gongfu session works in favour of this one, which is a neat illustration of how little sense it makes to talk about heat loss as good or bad in the abstract. It also explains why near-boiling water, which would be reckless with a western dose and a three-minute steep, is entirely reasonable here: the water is hot for the first few minutes and steadily less so afterwards, and the leaf load is too low for that early hot period to do any damage. The method is self-correcting in two independent ways at once, which is why it is so forgiving and why it has survived as everyday practice for so long.
What the leaf has to be like
Not every tea suits it, and the constraint is physical rather than about quality. Fine broken leaf and dust are unpleasant to drink around, since they do not sink and end up in the mouth, and they extract too fast for the plateau to establish gently. Very tightly rolled leaf takes so long to open in a glass that the first half of the session is thin and the second half arrives as a surprise. What works is intact, medium-to-large leaf that wets and sinks reasonably promptly: most Chinese green tea, yellow tea, coarse late-picked white tea, and lightly processed material generally. Teas whose whole appeal depends on a controlled short extraction — which is most good oolong and most fine black tea — are simply the wrong shape for the method, and give a flattened, averaged version of themselves that is not bad so much as beside the point.
The same mechanism, in other vessels
A vacuum flask is this arrangement with the cooling removed, which is exactly why flask tea is reliably harsher than glass tea from the same leaf and the same dose: the temperature stays high, so extraction never slows for that reason and only the concentration limit is holding it back. A bowl of tea drunk with the leaf still in it, as in older Chinese practice, is the same again with faster cooling and a wider surface. And a mug with a teabag left in is the identical physics at a dose high enough that the plateau lands well past drinkable — which is why that produces the bitter, stewed cup everybody recognises, and why the instruction to remove the bag is not fussiness. One mechanism, four outcomes, and the difference between them is almost entirely how much leaf is sitting in the water.