Different Time On and Off the Wrist? That Is Positional Variance

Daniel Hart · ·

A balance wheel seen at an angle above a movement plate
Gravity pulls on the balance differently depending on how the watch lies.

If your watch keeps good time on your wrist and drifts overnight on the nightstand, nothing is wrong with it. You have noticed positional variance, which every mechanical watch has.

Gravity pulls on the balance wheel differently depending on how the watch is lying, and the rate changes accordingly.

The useful part is that you can use this rather than merely suffer it.

The short answer

Why position changes the rate

The balance wheel swings on a pivot. Lying flat, the pivot bears the wheel’s weight on its end. Standing on edge, it bears the weight on its side, and the friction and the effective geometry are different.

Those differences are small and they add up over 24 hours. A watch can easily run five seconds a day faster crown-up than dial-up, and neither figure is a fault.

This is why chronometer certification is not a single measurement. The criteria published on the COSC site require testing across five positions and three temperatures precisely because one number would say very little.

Macro of a coiled balance spring above a movement plate
Positional variance is why chronometer testing uses five positions.

How to use it rather than suffer it

Measure the watch overnight in each of the common resting positions: dial up, dial down, crown up, crown left, crown down. One night each, five nights.

Then store it overnight in whichever position runs slowest. If the watch gains four seconds during the day on your wrist and loses three overnight dial-down, it is effectively keeping near-perfect time without a single adjustment.

Our accuracy calculator is the practical tool for this, because judging it by eye across five nights is unreliable.

Position Typical tendency
Dial up Often the fastest, and the flattest
Dial down Usually close to dial up
Crown up Frequently slower
Crown left Varies most between watches
Crown down Often the slowest
On the wrist A moving average of all of them
A watchmakers bench with tweezers and screwdrivers
It is adjustable, and how you store it is free.

When variance is too large

A spread of five to ten seconds a day between the best and worst positions is ordinary. A spread of thirty seconds is not, and it points at something worth looking at.

Large positional variance usually means the balance is out of poise, or the pivots are worn, and it is corrected during a service rather than by a regulation. A regulation shifts the whole rate; it does not narrow the spread between positions.

So the number to give a watchmaker is not just how fast it runs, but how much the rate differs between positions. That distinction tells them whether to adjust or to open it.

What this does not explain

If the watch suddenly gained minutes a day rather than seconds, this is not the cause. That is magnetism, and our running fast diagnostic separates it from a genuine fault.

If it is losing steadily in every position, that is energy rather than geometry. Our guide to a watch losing time covers the causes.

And if you own a common calibre, the expected figures are documented. Our NH35 explainer and ETA 2824-2 explainer cover two of the most widely used.

Frequently asked questions

Why does my watch keep different time on and off the wrist?

Positional variance. Gravity acts on the balance wheel differently depending on how the watch is lying, so the rate changes with position. Every mechanical watch does this.

How much positional variance is normal?

A spread of five to ten seconds a day between the best and worst positions is ordinary. Thirty seconds is not, and points at a balance out of poise or worn pivots.

Can I use positional variance to my advantage?

Yes. Find the position that runs slowest and store the watch that way overnight, cancelling out the gain it makes on your wrist during the day.

Why does chronometer testing use five positions?

Because a single measurement would say very little. COSC certification measures across five positions and three temperatures precisely because rate varies with both.

Will a regulation fix large positional variance?

No. A regulation shifts the overall rate but does not narrow the spread between positions. That needs the balance poising or the pivots addressing, which is service work.

The verdict

Positional variance is not a fault, it is physics, and the fact that you noticed it means you are measuring your watch properly.

Find the slow position and sleep the watch in it. That is a free correction that costs nothing and works on every mechanical watch.

Scroll to Top