Wind it once a day, at roughly the same time, and stop when you feel resistance — that is the routine I would pick over the "top it up whenever I remember" approach. The two methods people actually use are calendar winding (a fixed daily appointment, usually in the morning) and opportunistic winding (a few turns whenever the watch crosses your mind). Calendar winding keeps the mainspring in a narrower state of charge, which is easier on the movement and makes the watch's rate more consistent; opportunistic winding is more forgiving of a forgotten day but leaves you guessing how much power is actually stored. A mechanical watch is a timekeeper driven by a wound spring rather than a battery or quartz oscillator: energy stored in the mainspring is released through a gear train, metered by an escapement, and displayed by the hands. Everything below assumes that architecture, because the winding advice only makes sense once you know what you are feeding.
The two winding styles
Calendar winding means you turn the crown at a set point in the day — before you put the watch on, after you take it off, whenever it fits your habit — until the crown stops or the slip clutch begins to slip. In a hand-wound movement you will feel the spring reach full tension and the crown will simply refuse to go further. In an automatic, the crown usually keeps turning past full wind because a slipping mechanism protects the spring, so you are feeling for the point where it stops taking effort rather than for a hard stop. That difference matters: on an automatic, "it still turns" is not evidence that it needs more.
Opportunistic winding is what most owners drift into. You notice the watch has stopped, you give it a dozen turns, you wear it for two days, it stops again over a weekend. It works, but it produces two bad habits. The first is under-winding — a mainspring that never reaches full charge delivers less torque, and a movement running on low torque tends to lose amplitude, which shows up as a watch that drifts faster or slower than it does when fully wound. The second is over-eager winding on a hand-wound piece, where people keep pushing past the point of resistance because they have been told more power is better. It is not. Once the spring is fully coiled, additional force has nowhere to go except into the winding train.
If you own one mechanical watch and wear it daily, opportunistic winding will probably get you through. If you own several and rotate them, calendar winding is the only version that keeps you from discovering a dead watch on a morning you needed it.
How to wind it properly
Unscrew the crown if it is a screw-down type, then turn it clockwise with the watch off your wrist or on it — either is fine, but off the wrist gives you a better feel for the tension. Use the ball of your thumb and index finger rather than your fingertips, and turn in steady, unhurried strokes. Twenty to thirty full rotations of the crown is the usual range for bringing a drained automatic to full charge; a hand-wound movement takes fewer, and you should stop the moment you feel it firm up. Screw the crown back down if it came that way.
- Wind at the same time each day so you notice immediately when the watch feels different.
- Stop at resistance on a hand-wound movement — never force the last quarter turn.
- On an automatic, expect the crown to keep rotating past full wind; that is the slip clutch, not a sign you need more.
- Do not wind a watch you are about to store for months; let it run down instead.
- If the crown feels gritty, grinds, or the stem wobbles, stop and have it looked at rather than winding through it.
The one thing people get wrong most often is treating winding as a strength exercise. It is a feel exercise. If you cannot tell whether the spring is loaded, wind fewer turns more often until you learn the movement's signature.
Reading the power reserve
A fully wound mechanical watch typically runs somewhere between about a day and a week, depending on the movement, and the last stretch of that reserve is where timekeeping gets worst. Torque falls off as the spring unwinds, amplitude drops, and the rate wanders. This is why a watch that keeps excellent time on Monday can be noticeably off by Friday if you have not topped it up — the mechanism has not changed, only the energy feeding it.
The practical consequence: if accuracy matters to you, wind daily rather than waiting for the watch to stop. A watch that is always near full charge is a watch whose rate you can actually predict. If you only care that it is running and roughly right, letting it run down between wears costs you nothing but a reset.
There is a limit to how much this helps. A movement that is due for service — dried lubricants, a worn mainspring, a dirty escapement — will not be rescued by better winding habits. Full wind on a tired movement still gives you a tired movement. If a watch that used to run well starts losing time even when freshly wound, the answer is a service, not more turns of the crown.
When not to wind
Do not wind a watch that is going into storage for months. Let it run down, then put it away. A spring held at full tension for a long idle period is doing the movement no favours, and there is no benefit to keeping a watch you are not wearing in a state of readiness.
Do not wind through resistance on a hand-wound watch to "make sure". Do not wind a watch whose crown has been pulled out to a setting position — you will be turning the hands, not the spring, and on some movements you can damage the keyless works. Do not assume a watch that has been sitting in a drawer needs a full wind before you know whether it runs at all; give it a few turns, let it start, and see what it does.
Also worth saying plainly: winding a mechanical watch is not the same task as maintaining a mechanical keyboard, though both attract the same kind of owner. A mechanical keyboard is a keyboard whose keys each sit on an individual physical switch rather than on a rubber dome membrane, and cleaning one means pulling the keycaps, removing debris, and cleaning the switch plate and stems — a job with its own sequence and its own ways to go wrong. If you came here looking for that, the winding advice does not transfer.
A routine that holds up
Pick a time, wind to full, stop at resistance, and note how the watch behaves over the following week. That is the whole method. The reason it beats everything else is not that daily winding is inherently superior to any other interval — it is that a fixed habit gives you a consistent baseline. When the watch starts running fast or slow, you will know whether the cause is the winding or the movement, because the winding has not changed.
The trade-off is real: calendar winding demands a small daily act of attention, and it is easy to skip when you are travelling or wearing something else. If that describes you, accept opportunistic winding and just be honest that you are trading predictability for convenience. What you should not do is wind aggressively and irregularly, because that combination gives you neither the consistency of the first method nor the low effort of the second — only a watch whose state of charge you can never quite account for.