Problem
The hands do not line up with the markers
Short answer
Run one test first. Set the watch precisely, then look again after a day. If the offset is unchanged, the hand was fitted slightly off and this is cosmetic. If the offset has grown, and the hands no longer agree with each other, the cannon pinion is slipping and it needs a bench.
On this page (6 sections)
The display side of a mechanical watch is simpler than people expect. The center wheel turns once an hour. Over its arbor sits the cannon pinion, held only by friction, carrying the minute hand. A small gear train from the cannon pinion, the motion works, turns the hour wheel at one twelfth the speed, carrying the hour hand. The seconds hand presses onto a separate pivot from the fourth wheel.
Every one of those is a friction fit. None of them is keyed, splined or indexed. The position of a hand relative to a marker is entirely the product of where a human being pressed it on. That fact explains most of what owners report on this subject.
The test that settles it
Before diagnosing anything, run this. It takes a day and it separates the cosmetic case from the mechanical one.
- Wind the watch fully and set the minute hand exactly on a marker, with the watch running.
- Note the offset you can see between hour hand and hour marker, and between minute hand and minute track.
- Leave the watch running for 24 hours without setting it.
- Look again.
If the offsets are identical to what you noted, the hands are fitted slightly off and nothing is wrong mechanically. If the minute hand has fallen further behind, and has fallen behind relative to the hour hand as well as relative to real time, the cannon pinion is slipping. Those two results lead to completely different conversations and completely different bills.
A hand fitted slightly off
This is the common case, and it is a finishing issue rather than a fault. Hands are pressed on by an assembler or a watchmaker, sighted by eye, at a bench, under magnification, at a rate of many watches an hour. A degree or two of rotation on a minute hand is a fraction of a minute of indicated time and is well within the assembly tolerance of nearly every watch made.
Two things follow from that. First, it does not affect timekeeping in any measurable way: a minute hand a degree off reads about ten seconds out, which is smaller than the daily variation of most mechanical watches described in what counts as good accuracy. Second, it will come back. A watch that has been serviced has had its hands removed and refitted, and the new alignment is whatever the last watchmaker's eye produced.
Some brands treat hand alignment as part of what they sell. Grand Seiko is the usual example: the seconds hand aligned to the index is a stated part of the finishing, and it is one of the things the price buys, as our page on Grand Seiko sets out. Expecting that standard from a watch that was never built to it is a category error rather than a complaint.
The seconds hand and the index
The most frequent version of this question, and the one with the least alarming answer. A seconds hand that lands between marks is normal. There is nothing in a standard movement that positions it: it goes onto the fourth wheel pivot wherever the tool leaves it, and the dial or chapter ring printing has an independent tolerance.
On a watch with hacking seconds, you can at least stop and restart the hand at a chosen point, which is why hacking watches let you synchronize to a time signal. That does not align the hand to the printed marks. It aligns it to the correct time, which is the thing that actually matters.
A slipping cannon pinion
This is the real fault, and it disguises itself as poor timekeeping.
The cannon pinion is deliberately a friction fit. That friction is what lets you set the time: turning the crown drives the motion works and slips the cannon pinion against the center wheel arbor without disturbing the running train. It is engineered to have exactly enough grip to drive the minute hand and exactly little enough to yield under setting torque.
When wear, dry lubricant or a previous bad adjustment reduces that grip, the minute hand starts falling behind the movement. Symptoms:
- The watch appears to lose time, sometimes badly, but a timegrapher reading shows the movement is running perfectly on rate.
- The minute hand and hour hand stop agreeing with each other over days.
- Setting the time feels loose, with no resistance from the crown in the setting position.
- In severe cases the minute hand can be nudged by a knock and stay where you put it.
That first symptom is the important one. Rate testing looks at the balance, not at the hands, which is exactly why an owner can pay to have a watch regulated twice and see no improvement. If you have been chasing an accuracy problem without success, this is worth ruling out before anything else, and our guide to timing your own watch explains how to compare movement rate against displayed time properly.
The fix is a bench job. The cannon pinion is either tightened, by lightly pinching the waisted section with a purpose-made tool, or replaced. It is usually done as part of a service rather than alone, because the movement has to come out of the case and the hands have to come off either way.
When it is the chapter ring, not the hands
On many dive watches and sports watches the minute track is printed on a separate sloping ring that sits between the dial and the crystal. That ring is located by the case and the movement holder, and it can be fitted a degree or two rotated. The result: hands that agree with each other perfectly, and a minute track that is visibly off at twelve.
This is a case assembly matter. It is corrected by opening the watch and re-seating the ring, not by touching the movement or the hands, and it is one of the few alignment complaints that is genuinely a manufacturing defect worth raising under warranty. It is well documented on several popular divers, which is why it is worth checking specifically when buying, as covered in buying pre-owned.
What is worth correcting
A slipping cannon pinion: always. It costs you real timekeeping, it gets worse, and in the worst case the hands stop tracking the movement entirely. Budget $100 to $250 for the adjustment alone at an independent watchmaker, or fold it into a full service, which for a common Swiss three-hander is typically $250 to $500 at an independent and $350 to $600 at a brand service center. Figures are indicative US dollars for 2026, quoted before sales tax, and they vary by country, by parts channel and by whether the movement is serviceable at all. Our servicing cost page goes through the variables.
A chapter ring rotated on a new watch: raise it with the retailer while you can. It is a defect and it is cheap to correct.
A hand fitted a degree off: usually not. The correction costs a bench visit, the outcome depends on the eye of whoever does it, and every hand removal is a chance to scratch a dial. Ask for it at the next service, when the hands are coming off anyway, and accept that the result is a human judgment rather than a specification. Legibility, not perfection, is what dials are actually designed around, as our guide to dials, hands and legibility explains.
Common follow-up questions
My seconds hand never hits the marks. Is my watch badly made?
Almost certainly not. Nothing in a standard movement synchronizes the seconds hand to the printed index. It is pressed onto a pivot in whatever rotational position the hand-setting tool leaves it, and the dial printing has its own tolerance on top. A few makers, Grand Seiko among the best known, make a point of aligning it precisely, and they charge for the labor that takes. On everything else it is normal.
How do I know it is the cannon pinion and not the movement running slow?
Compare the hands against each other, not just against the correct time. If the movement is running slow, all the hands stay in agreement and the whole display drifts together. If the cannon pinion is slipping, the minute hand falls behind while the hour hand and seconds hand keep their relationship to the movement, so the hands stop agreeing with one another. That disagreement is the diagnostic.
The hour hand is not quite on the hour when the date flips.
Two possibilities. If the date changes near midnight but the hour hand reads a little before or after twelve, the hour hand was fitted slightly off and it is cosmetic. If the date changes at what your watch calls noon, the hour hand is fitted twelve hours out, or more likely you set the time through the wrong twelve-hour cycle. That second one is a setting habit, not a fault, and it is covered on our date change page.
Can I push a hand back into position myself?
Strongly discouraged. Removing a hand needs a purpose-made lever or presto tool, a dial protector, and hands that are supported so they do not bend. Refitting needs the right diameter pusher and enough force to seat the hand without pushing the cannon pinion down onto the dial. The usual amateur outcome is a scratched dial, a kinked hand, or a hand refitted at a height where it fouls the one above it.
Sources and further reading
- Standard watchmaking bench practice on motion works, which describes the cannon pinion as a friction fit on the center wheel arbor, sized to slip during hand setting and to drive the minute hand at all other times.
- ETA and Sellita technical communications for the 2824-2 and SW200-1, which detail the motion works, the hand fitting heights and the cannon pinion adjustment.
- Seiko Instruments technical guides for the NH35 and 4R series, which specify hand fitting order and the clearances between hour, minute and seconds hands.
- Grand Seiko product documentation describing the finishing and hand-fitting standards applied to its watches, including alignment of the seconds hand to the index.
- ISO 3159, Horology: Wrist-chronometers with spring balance oscillator, which defines rate testing on the movement rather than on the displayed time, which is why a slipping cannon pinion is invisible to a rate test.
Last reviewed 4 September 2026. Spotted an error? Tell us and we will fix it in public.