The clock nobody agrees on
A stated steep time is a duration between two events, and brewing instructions almost never say what those two events are. Does the clock start when the water begins going in, or when the vessel is full? Does it stop when you begin pouring, when the last drop has left, or when the cup reaches the table? Different sources assume different answers without ever declaring them, which means the same printed number describes materially different steeps depending on who wrote it and what they were brewing in. For a western three-minute brew the ambiguity amounts to a few per cent and can be ignored entirely. For a ten-second gongfu steep, the pour in and the drain out can together exceed the stated time, and ignoring them is a large part of why two people following one published schedule end up with two quite different sessions and each conclude the other is doing it wrong.
Pouring in is already steeping
The moment water touches leaf, extraction begins, and the leaf that was wetted first has been extracting for the whole time it takes to fill the vessel. Filling a hundred-and-fifty-millilitre gaiwan takes several seconds, and filling it slowly in a spiral around the edge takes appreciably longer than filling it fast in the centre — so pour speed is quietly a brewing variable rather than a matter of style. It also changes how much the leaf bed is agitated and how it settles, which affects how evenly the water reaches it. The consistent practice is to fill at the same speed every time, in the same pattern, and to start the clock when the water is in, accepting the fill as part of the steep and holding it constant rather than trying to exclude it from the accounting. Consistency is worth more here than precision, because precision is not available.
Draining is steeping too
The same argument applies at the other end, and it matters more, because the vessel is full and the liquor is at its most concentrated while the pour is happening. A pot that takes fifteen seconds to empty has added fifteen seconds of extraction to the steep, distributed unevenly — the last of the liquid has been in contact substantially longer than the first and has been passing through a leaf bed that is more concentrated with every second. This is where vessel choice and brewing time interact directly rather than as separate topics. A slow-draining pot simply cannot be used for very short steeps at all, because the drain alone exceeds the intended contact time, and no amount of care with a timer will make it possible. That constraint decides what a vessel can be used for more firmly than its material does.
Why this bites hardest at the short end
The proportional error is what matters rather than the absolute one. Five seconds added to a three-minute steep is under three per cent and completely invisible in the cup. Five seconds added to an eight-second steep is more than sixty per cent, and is the difference between a bright, clean, aromatic cup and a thick and slightly gripping one. This arithmetic is the whole explanation for the observation that gongfu brewing is fussy: it is not that its practitioners are precious about small things, it is that at that time scale the mechanics of pouring stop being a rounding error and become a first-order variable, on a par with temperature. It is also why gongfu schedules are written on the assumption of vessels that fill and empty in a couple of seconds, and why they transfer so badly to a pot that does not.
Flash steeping, and what zero seconds means
A flash steep is a pour in and a pour straight back out with no dwell at all, and the tea is nonetheless extracted — because the fill and the drain are themselves several seconds of contact, most of it at full temperature with fully open leaf and a very high leaf-to-water ratio. It is used at the start of a session where even a nominal ten seconds would already be too much, and with material that gives extremely fast: fine broken leaf, dust grades, a heavily broken cake, or any session running at the top of the ratio range. Reading a flash steep as no time at all misses what is happening. It is the shortest steep the mechanics of the vessel allow, the vessel is what sets its actual length, and two different pots will produce two different flash steeps from the same leaf.
A convention worth adopting
Start the clock when the vessel is full, stop it when you begin to pour, and choose a vessel that drains fast enough that the difference between that and any other convention is small. Then keep it, without exception. The value is not that this particular convention is correct — none of them is, and the choice is arbitrary — but that holding one constant makes your own recorded times comparable across teas, sessions and years, which is the only circumstance in which a written-down time means anything later. Every schedule in this catalogue is written on that convention. It is stated here because it is otherwise entirely invisible to a reader, and because a reader using a different one will find the numbers systematically a little short.
When the clock stops mattering
In grandpa brewing, in cold brewing and in any boiled preparation, there is no meaningful start and stop at all, and the numbers recorded for those methods are conventions — when to take the first drink, how long to hold a simmer — rather than measurements of contact time. Trying to time them precisely is applying the wrong model and produces false confidence rather than better tea. The general principle is that a clock is genuinely useful exactly where the extraction is being deliberately interrupted and the interruption is the technique, and it is decorative everywhere else. Recognising which of the two situations you are in is worth more than any timer, and it also explains why some very good tea drinkers appear never to time anything: they are brewing in the styles where timing does not decide the outcome.