Watch Running In Only Certain Positions: Watch Only Runs Correctly in Certain Positions? Here’s Why
If you’re reading this, you’ve probably already done something most watch owners never bother with: you’ve actually tracked how your watch runs in different positions. You laid it flat overnight, checked the rate in the morning, then tried it crown-up the next night, maybe crown-down after that. And now you’re staring at a spreadsheet (or a mental note) that doesn’t quite add up. Your watch gains fifteen seconds dial-up but loses five seconds crown-down. Or it’s dead accurate on the nightstand but loses half a minute whenever you wear it actively.
That feeling of ‘wait, is this normal?’ is exactly what this guide addresses. Positional error — the technical term for a watch that runs differently depending on how it’s oriented — is one of the most common discoveries for anyone who starts measuring accuracy seriously. The good news is that some positional variance is completely normal, even in expensive watches. The trick is knowing where ‘normal’ ends and where a real mechanical issue begins.
What Positional Error Actually Is
At its core, positional error comes down to one thing: gravity pulling on the balance wheel differently depending on which way the watch is facing. Your watch’s balance wheel oscillates back and forth, and its rate depends on the hairspring’s ability to return it to center consistently. But the balance staff rides on pivots that sit in jewel bearings, and those pivots have microscopic clearance — we’re talking thousandths of a millimeter. When you rotate the watch, gravity shifts the balance staff slightly off-center within those clearances.
That tiny shift changes the friction profile on the pivots and subtly alters the effective length of the hairspring. The result is a measurable change in rate: the watch might run faster in one position and slower in another. This is not a defect. Every mechanical watch with a traditional Swiss lever escapement exhibits some positional error. It’s a fundamental consequence of how mechanical timekeeping works in a gravity field.
Good to know:
This is exactly why marine chronometers and high-end pocket watches were historically adjusted to multiple positions — five positions was the gold standard for observatory-grade instruments. A modern ETA 2824 or Sellita SW200 is typically adjusted to three positions at the factory. The more positions a movement is adjusted to, the smaller the rate difference between them.
Normal Positional Variance vs Excessive Variance
So what’s actually normal? This is where a lot of hobbyists get nervous unnecessarily. A watch that gains 8 seconds dial-up and loses 4 seconds crown-down might look alarming on paper, but in practice it’s well within what most watchmakers consider acceptable for a standard-grade movement. The industry benchmark for positional error — the maximum allowed difference between any two positions — varies by movement grade, but here’s a realistic breakdown of what you should expect:
| Movement Grade | Typical Positional Variance (max spread) | Example Movements |
|---|---|---|
| Standard / Unadjusted | 15–25 seconds per day | NH35, Miyota 8215, basic ETA 2824-2 |
| Elaboré / Adjusted to 3 positions | 8–15 seconds per day | ETA 2824-2 (Elaboré), Sellita SW200-1 |
| Top / Adjusted to 5 positions | 4–8 seconds per day | ETA 2892-A2 (Top), Valjoux 7750 (Top) |
| Chronometer-grade (COSC) | −4 to +6 overall; positional spread typically 3–6s | Rolex 3135, Omega 8500, ETA 2892-A2 (Chronometer) |
The key number here is the spread — the difference between your best and worst positions. If your watch runs +12 in one position and −5 in another, that’s a 17-second spread. For a standard movement that’s not great but not shocking. For a chronometer-grade movement, that would be a red flag worth investigating.
But here’s the nuance that most articles miss: positional variance matters most when it affects your real-world daily rate. If your watch runs fast dial-up but you never store it dial-up, who cares? The practical question is whether the positions you actually use — wearing it on your wrist (which is mostly dial-up or crown-up depending on how you hold your arm) and storing it at night — give you a consistent overall rate you can live with.
Tip:
If your positional spread is under 15 seconds and your watch’s average daily rate across all positions is within ±10 seconds, you probably don’t have a problem — you just have a normal mechanical watch. Try storing it overnight in the position that offsets your daytime rate. For example, if it gains during wear, store it crown-up (usually the slowest position) to balance out.
How Watchmakers Test and Record Positional Error
When a watchmaker services your movement, they don’t just wind it up and call it done. They run a multi-position timing test using a timing machine — typically a Witschi or a Chinese clone — that measures the watch’s rate, amplitude, and beat error in six standard positions. These positions are dial-up (DU), dial-down (DD), crown-up (CU), crown-down (CD), crown-left (CL), and crown-right (CR). The watch runs in each position for 30 to 60 seconds, and the machine records the instantaneous rate.
The reason they test all six positions isn’t because your watch will spend equal time in each one. It’s because the combination of results tells the watchmaker where the problem is. A watch that shows a large rate difference between dial-up and dial-down, for instance, often points to a poise issue on the balance wheel. A watch that runs fast in crown positions but normal in dial positions might indicate worn pivot holes on the balance staff or a hairspring that’s slightly out of flat.
If you’re doing this at home with a phone app timing tool (like Watch Accuracy Meter or a similar app), you’re getting a rough approximation — useful for trend-spotting, but not precise enough to diagnose a specific fault. The microphone on your phone picks up the tick sound and estimates the rate, but it can’t measure amplitude or beat error, which are critical for understanding why the positional error exists.
Caution:
Don’t try to regulate your watch based solely on phone app readings in different positions. Phone apps can’t measure amplitude, and low amplitude can create misleading positional error readings. If you see a spread over 20 seconds on an app, take it to a watchmaker who can put it on a real timing machine — the app might be showing a symptom of a deeper issue that regulation won’t fix.
Common Causes of Excessive Positional Error
Once you’ve confirmed that your positional spread is genuinely excessive — say, over 20 seconds for a modern automatic or over 30 seconds for an older manual-wind — it’s time to think about what’s causing it. There are four main culprits, and they range from easy fixes to full-service issues.
Worn Pivots and Jewel Holes
This is the most common cause in older watches or watches that haven’t been serviced in 7–10 years. The balance staff pivots are tiny — on a typical ETA 2824 they’re about 0.08mm in diameter. Over years of running, those pivots wear against the jewel holes, creating an oval-shaped clearance. When you rotate the watch, the balance staff shifts more than it should within that oval, causing a much larger rate change than when the pivots were round and tight.
This is a full-service issue. You can’t fix worn pivots without replacing the balance staff (or the entire balance assembly) and potentially the jewel settings. If you’re hearing a grinding or gritty feeling when you hand-wind, or if the amplitude is below 200 degrees in any position, worn pivots are almost certainly the culprit.
Balance Wheel Poise Issues
A balance wheel should be perfectly balanced — meaning its center of gravity coincides exactly with its axis of rotation. If it’s even microscopically heavy on one side, gravity will pull that heavy side down in different positions, altering the rate. This is called a ‘poise error.’ It shows up as a consistent pattern: the watch runs faster in dial-up than dial-down, or vice versa, with the crown positions showing less difference.
Poise issues can sometimes be corrected by a skilled watchmaker without replacing parts. They use a poising tool (essentially a set of parallel knife-edges) to find the heavy spot, then carefully remove material from the balance wheel rim using a small file or a carbide cutter. This is not a home repair — one wrong stroke and you’ve ruined the balance wheel.
Magnetization
Magnetization is the wild card because it can create positional error that looks like a poise issue or a pivot problem. When a hairspring becomes magnetized, its coils stick together in certain positions, effectively shortening the spring and making the watch run fast. But the sticking is position-dependent — it might happen dial-up but not crown-down, because gravity helps or hinders the coils from separating.
The giveaway is that magnetization usually causes the watch to run consistently fast in all positions, but with a larger variance than normal. If your watch is gaining 20 seconds dial-up and 30 seconds crown-down, magnetization is a strong possibility. It’s also the easiest fix — a demagnetizer costs about $20 and takes ten seconds. But be careful: cheap demagnetizers can sometimes magnetize a watch further if used incorrectly.
Tip:
You can check for magnetization at home with a simple compass. Hold the watch near a compass needle — if the needle deflects noticeably when you move the watch around, you’ve got magnetization. Just remember that the watch itself contains magnetic parts (the mainspring barrel, some steel screws), so a tiny deflection is normal. If the needle swings 30 degrees or more, demagnetize the watch.
Hairspring Issues (Out of Flat, Out of Center)
The hairspring is the most delicate component in the movement. If it’s been knocked (say, from a drop or a sharp impact), it can become ‘out of flat’ — meaning it doesn’t sit perfectly parallel to the balance wheel — or ‘out of center,’ meaning the coils aren’t concentric. Both conditions create positional error because the spring’s geometry changes as gravity shifts the watch.
This is almost always a watchmaker job. Hairspring adjustment requires a steady hand, a loupe, and specialized tweezers. Attempting to bend a hairspring back into shape at home is a fast track to destroying it. If you suspect a hairspring issue — especially after a drop — get it serviced.
What a Regulation/Adjustment Can and Can’t Fix Here
This is where a lot of hobbyists get tripped up. Regulation — moving the regulator pins to speed up or slow down the watch — changes the average rate across all positions. It does not correct the spread between positions. If your watch runs +10 dial-up and −5 crown-down, regulating it to run +2 dial-up will also shift the crown-down rate to −13. You’ve reduced the overall average but made the spread worse in absolute terms.
True adjustment for positional error involves the things we discussed above: correcting poise, addressing hairspring geometry, and ensuring the balance staff pivots are round and the jewel holes are clean. That’s a service-level intervention, not a quick regulator tweak. A good watchmaker can reduce a 25-second spread to 8 seconds, but it takes time and skill.
What regulation can do is optimize your watch for how you actually wear it. If you know your watch runs +8 during daily wear (which is mostly dial-up and crown-up positions) and you store it crown-down (which runs −4), a regulator adjustment that brings the dial-up rate down to +4 will give you a net daily rate close to zero when you factor in overnight storage. That’s not fixing the positional error — it’s working around it.
Caution:
If you regulate a watch that has low amplitude (under 220 degrees) or a significant beat error (over 1.0ms), you’re putting a band-aid on a bullet wound. Low amplitude causes inconsistent rate in all positions, and regulating will only mask the symptom. Always check amplitude first — if it’s low, the watch needs service, not regulation.
When This Is Worth Addressing vs Living With
Here’s the honest answer that most guides won’t give you: for 80% of watch owners, positional error is something to be aware of but not something to fix. If your watch has a 15-second spread between positions but you’re getting ±10 seconds per day in real-world wear, just live with it. Store it in the position that offsets your daily rate and move on. The cost of a full service to reduce positional spread from 15 seconds to 5 seconds is usually $300–$600, and that money buys a lot of other things.
When should you actually address it? Three situations:
- Your positional spread exceeds 25 seconds and your daily rate is consistently over 20 seconds off, even after trying different storage positions.
- You’ve noticed the positional error getting worse over time — that’s a sign of progressive wear on pivots or jewels, and it won’t get better on its own.
- You’re chasing chronometer-level accuracy and the spread is preventing you from hitting −4/+6 in any reasonable wearing scenario.
For the enthusiast who’s deep enough to be reading this guide, the real value isn’t in fixing every positional quirk. It’s in understanding what your watch is telling you. A watch that used to have a 10-second spread and now has a 30-second spread is trying to tell you it needs service. A watch that has always had a 15-second spread is just being itself. Learn the difference, and you’ll save yourself a lot of worry — and a lot of money.
Positional error is not a flaw. It’s a feature of mechanical watchmaking that you now understand well enough to interpret. Use that knowledge to make better decisions about when to service, when to regulate, and when to just set the watch down crown-up and enjoy the ticking.
Frequently Asked Questions
Is some positional error normal even in a well-running watch?
Absolutely. Even a freshly serviced, well-regulated watch will show a few seconds of variation between positions due to gravity’s effect on the balance wheel and hairspring. A spread of 5 to 10 seconds across six positions is perfectly normal for most movements, and many COSC-certified chronometers still show 4 to 6 seconds of positional deviation. The goal isn’t zero spread — it’s a spread that’s consistent and predictable enough for you to work around.
Can positional error be fixed through regulation alone?
Regulation adjusts the overall rate — how fast or slow the watch runs on average — but it cannot fix the spread between positions. If your watch runs +10 seconds dial-up and −15 seconds crown-down, moving the regulator will shift both numbers equally, leaving the 25-second spread intact. Addressing positional error usually requires dynamic poising of the balance wheel, adjusting endshake, or correcting hairspring centering — work best left to a watchmaker with a timing machine.
How do watchmakers actually test positional accuracy?
They use a timing machine that measures rate, amplitude, and beat error in each of the six standard positions: dial-up, dial-down, crown-up, crown-down, crown-left, and crown-right. Each position is measured for 30 to 60 seconds after the watch has settled in that orientation, and the results are recorded as a positional error chart. This chart tells the watchmaker not just the spread, but which specific positions are weak, helping them diagnose issues like a bent pivot or an out-of-poise balance.
🛒 Shop Compatible Parts
Prices and stock verified daily by our affiliate partners. As an affiliate, WatchPartsFit may earn a commission on qualifying purchases at no extra cost to you.
Check Compatibility First →