How long does a mechanical watch last?

A mechanical watch has no service life in the engineering sense. There is no component inside it that is expected to fail at a known age, and there is nothing that cannot be replaced by a competent watchmaker. Movements built in the 1890s are running today, keeping time to within a minute a day, on their fourth or fifth mainspring and their second balance staff.

So the honest answer is: as long as you are willing to pay for it, and as long as somebody still makes the parts. The first half is a decision. The second half is out of your hands, and it is where watches actually die.

What a movement is actually doing

The scale of the work helps explain why servicing exists at all. A balance wheel running at 28,800 vibrations per hour swings eight times a second: about 691,200 vibrations a day, roughly 252 million a year. Each one is a controlled collision. The escapement locks, releases, delivers an impulse and locks again, and the escape wheel itself turns something like ten to sixteen times a minute depending on the beat rate.

Every one of those events happens on a film of oil a few microns thick. The steel and the ruby will outlast everyone reading this. The oil will not, and that is the whole basis of servicing. It thins, spreads away from where it was placed by capillary action, oxidizes and eventually dries to a varnish that increases friction rather than reducing it. Run a movement with dry pivots and you convert a cheap oil change into a parts replacement.

What wears out first

In order of how often a watchmaker meets them:

• Pivots. The ends of the wheel arbors, particularly the balance staff, are roughly a tenth of a millimeter across. Running dry, a pivot polishes down and goes out of round, and the jewel hole it runs in wears oval. This is the classic consequence of a long-postponed service.
• The mainspring. It fatigues, and eventually it breaks, usually at the eye or at the bridle. It is a consumable, and it is cheap.
• The barrel wall and arbor. In an automatic the mainspring's bridle deliberately slips against the barrel wall at full wind, on a special braking grease. When that grease ages, the slip point drifts and the wall polishes.
• The cannon pinion. A friction fit that lets you set the hands. It loosens with age, and the symptom is hands that fall behind or refuse to move.
• Reversing wheels and the rotor bearing. The rotor is a heavy weight on a small bearing, and it is the part most often damaged by people shaking watches to start them.
• Gaskets. Rubber has a genuine shelf life. This is why pressure testing belongs with every service.

Every item on that list is a stock part on a common caliber. None of them is a reason a watch stops existing.

The mainspring and the barrel

The mainspring deserves its own note because its history explains why old watches often run worse than they should. Springs made from carbon steel took a permanent set: they gradually lost their curvature and their torque, so a fifty-year-old watch on its original spring often shows low amplitude and a short power reserve even after a clean and oil. Modern springs use cobalt-nickel alloys that hold their form almost indefinitely, and fitting one to an older watch typically restores the reserve to its published figure.

The barrel is more interesting. On a hand-wound watch the spring is hooked to the wall and nothing rubs. On an automatic it is designed to slip, thousands of times over the watch's life, which means the barrel wall is a wear surface by intent. Barrels are usually replaced as a complete assembly rather than repaired, and on a common caliber that is a routine part order.

What almost never wears

The parts most owners worry about are the ones least likely to be a problem. A jewel is synthetic corundum at Mohs hardness 9, running against steel at Mohs 4 to 5. In a properly lubricated watch it effectively does not wear. Jewels are replaced because they crack under a shock, because they have worked loose in their setting, or because a dry pivot chewed the oil sink, and almost never because they simply wore out.

The plates, bridges and wheels are in a similar position. Absent corrosion, they are usually reusable indefinitely. A watchmaker inspecting a hundred-year-old movement is looking at pivots, the spring, the barrel and the state of any threads, not wondering whether the mainplate is tired.

What actually kills a watch mechanically

When a movement really is beyond economic repair, it is usually one of four things, and only one of them is wear.

Note what is absent from that table: old age. A movement does not become unserviceable through the passage of time. It becomes unserviceable because something happened to it.

Repairability is a commercial question

Here is the part that decides whether your watch is a hundred-year object or a twenty-year one. A watchmaker can replace a mainspring, a staff or a barrel only if there is a part to fit. Three things determine that.

How many were made. A caliber produced in the millions has an aftermarket. Generic mainsprings, staffs and stems exist for old ebauches, and donor movements are cheap. A caliber produced in the low thousands has neither.

Whether the maker still sells parts. Some brands sell parts to any qualified watchmaker. Several large groups supply them only to their own service centers and authorized partners, a policy that has been examined by the Swiss competition authority in the context of movement supply and by the European Commission after a complaint from the European repairers' association. The legal picture differs by market, but the practical picture is simple: a restricted-parts watch is repairable only for as long as the brand chooses to repair it.

Whether the brand still exists. Defunct brands with proprietary calibers are the hardest case in vintage watchmaking, which is why the vintage buying guide treats caliber identity as more important than the name on the dial.

Which watches will still be running in 2100

Two watches can be identical in quality and have completely different life expectancies.

The strong case: a movement made in enormous numbers, still in production or recently so, with published service documentation and an open parts supply. The ETA 2824 and 2892 families, the Valjoux 7750, Seiko's 7S and NH series and Miyota's 8 and 9 series all qualify. So do most mid-century Swiss ebauches, because of the sheer number of donors.

The weak case: a low-volume proprietary caliber from a brand that no longer exists, or one whose parts are locked inside a service network. It will run beautifully until the day it needs something specific, and then it becomes a problem for a specialist with a lathe.

Between those poles, a watch from a large brand with a long institutional memory is a reasonable bet. Some maisons publicly commit to restoring anything they have ever made, and that commitment, not the metallurgy, is the thing worth paying for if longevity is your goal.

What to do if you want yours to last

1. Service it on symptoms, with the published interval as a backstop. Falling reserve, a rate you cannot regulate out, or a gritty crown means now, whatever the calendar says. The reasoning is in how often a mechanical watch should be serviced.
2. Budget for it before you buy. Run the numbers through the cost of ownership calculator, and read what a service costs. A cheap watch with an expensive movement is a trap.
3. Keep water out. Corrosion is the one failure mode that regularly writes a movement off, so have the seals tested rather than trusting the number on the dial.
4. Keep the papers, the spare links and any original parts. They matter for value and they help the next watchmaker.
5. If you are buying specifically for the long term, buy the common caliber. It is the least glamorous advice on this site and the most reliable.

If you want an object your grandchild can wear, the decision is not about hardness, water resistance or finish. It is about whether a watchmaker in 2075 can order a mainspring for it. Ask that question before you buy, because it is the only one you cannot fix later.