Positional Error and Why It Isn't a Flaw

Same watch, same mainspring, same twenty minutes on the timegrapher — and four different numbers. Dial up, it reads +3 seconds a day. Set the case flat the other way, dial down, and it settles at +1. Stand it on its edge, crown pointing at the ceiling, and it climbs to +9. Nothing inside the movement changed between those readings except the orientation of the case itself, and a beginner's first instinct is to treat that spread as evidence of a bad regulation, a rushed service, or a balance that can't make up its mind. It's none of those things. A movement's rate isn't one fixed number that positional testing occasionally fails to confirm — it's a small family of numbers, one per orientation, and the differences between them are gravity doing exactly what gravity does to a rotating mass on a pivot.

Why Orientation Changes the Rate

A balance wheel sits on a staff no thicker than a few strands of hair, and that staff ends in two conical pivots resting in jeweled bearings — one under the balance cock, one in the mainplate. Lay the movement dial up and the weight of the balance assembly presses straight down through the lower pivot onto its cap jewel, a small, nearly frictionless point of contact. Turn the same movement so the balance staff runs horizontally instead — crown up or crown down — and gravity no longer pulls along the length of the staff. It pulls sideways, across it, so the pivot now bears against the wall of its jewel hole rather than the tip of a cap jewel. Friction at a flat point of contact isn't the same as friction at a rounded one, and that small difference in drag is enough to change how much energy actually reaches the balance on every swing, which is precisely what a rate reading measures.

The Six Standard Positions

A full positional check works through six orientations, not two, because a wristwatch spends its life in all of them — pocket, wrist, nightstand, the crown pointing wherever it was last set down. Watchmakers shorthand them DU, DD, CU, CD, CL, and CR: dial up, dial down, and crown up, down, left, or right for the four ways a case can stand on its edge. Amplitude tends to drop a little as the case moves from flat to vertical — see Reading Amplitude on a Timegrapher for what that arc looks like on its own — and rate moves with it, position by position, in a pattern that repeats caliber to caliber more than it varies watch to watch.

What a Normal Spread Looks Like

Talk of "positional error" as if it were a single number simplifies past the point of usefulness. What actually gets reported is a spread — the difference between the best and worst rate across the six orientations — and for a mechanical wristwatch caliber from any era, a spread of ten to twenty seconds a day between dial-up and a vertical position is common on a perfectly sound movement. Older calibers with plainer pivots, or movements running lower amplitude to begin with, tend to sit at the wider end of that range. None of it means the escapement is broken. It means gravity has more leverage on the balance in some orientations than in others, and the balance is reporting that leverage honestly, beat after beat.

A balance wheel has no opinion about which way is up. It only reports how hard gravity is working against whichever pivot happens to be carrying its weight at that moment.

Reading a Position Report Without Alarm

A position check earns its keep when the six numbers are read as a pattern rather than as six isolated grades. A normal-looking report tends to share a few traits:

• Dial-up and dial-down usually sit closest together, since both load a cap jewel rather than a jewel-hole wall.
• The four vertical positions cluster within a similar range of each other, even if that range sits well away from the two flat readings.
• The spread holds roughly steady across two or three separate winds of the mainspring, rather than growing wider each time.
• Amplitude falls a little in the vertical positions along with rate, moving in the same direction, not against it.

A watch that fits that pattern is behaving the way its caliber was designed to behave, even when the printed spread looks larger than a novice expects. The mainspring's own torque curve compounds this in the first hours after a wind, briefly raising amplitude and, with it, positional sensitivity.

When the Spread Does Mean Something

What deserves a second look isn't the existence of positional error — it's a spread that breaks the pattern above. A crown-down reading dramatically worse than crown-up, crown-left, and crown-right combined, rather than sitting near them, often points to a single damaged pivot rather than ordinary gravity loading. A dial-up/dial-down difference far larger than the vertical spread can indicate a balance that isn't poised: its mass isn't distributed evenly around the staff, so gravity finds a heavy side to pull on even flat on the bench. Dry or contaminated oil at one jewel and not its neighbor shows up the same way — fine in the position that doesn't load that pivot, noticeably worse in the one that does. None of these are things a spread alone proves; they're things an unusual spread is worth investigating, in much the same way a caliber's pivots eventually earn the lubrication check described in The Case for Re-Oiling Every Five Years.