Phone GPS accuracy

Two errors, not one

The frustration is universal and the explanation is boring: there are two independent sources of error and they add up.

The first is yours. A consumer satellite fix under open sky is accurate to a few metres. Under tree canopy, beside a stone wall, in a narrow valley or among buildings it is worse, because the signals reaching you have bounced off something before arriving and the receiver cannot always tell.

The second is the owner’s. The coordinates printed on the listing were taken by another person, with a device subject to exactly the same physics, standing in the same awkward place, possibly in a hurry, possibly years ago, possibly averaged carefully and possibly not.

When both errors happen to point the same way, you can be ten metres from a container with every piece of equipment working exactly as designed.

What actually improves your fix

Four things help, in roughly this order.

Multi-band hardware. A receiver that listens on more than one satellite frequency can discard some reflected signals. Where a phone has it, the difference under canopy and beside buildings is visible rather than theoretical. It is the only hardware specification on this page worth shopping for.

Standing still. A fix improves for the first half-minute or so as the receiver accumulates observations. Arriving, stopping, and letting it settle before searching costs thirty seconds and routinely saves ten minutes.

Walking a circle. Rather than trusting one reading, walk a slow loop of ten or fifteen metres around the area and watch where the arrow consistently points back to. The centre of that behaviour is a better estimate than any instantaneous number.

Network assistance. Letting the phone use mobile and wireless signals alongside satellites helps where those networks are dense, which happens to be where satellite reception is worst.

What does not help

Restarting the app. It feels productive. It does not change the physics.

A more expensive phone, unless the expense bought multi-band. Price and accuracy are only loosely related.

Trusting the accuracy figure on screen. It is the device’s own estimate of its own error, which is itself an estimate. It is a useful relative signal, worse in the woods than in the car park — and a poor absolute one.

When it is the listing, not you

A large share of “my phone is wrong” turns out to be “the coordinates are wrong”, and the evidence is usually already on the page.

Read the recent logs before you decide your device is at fault. If several independent finders mention the coordinates being consistently off in the same direction, that is a much stronger measurement than your one visit: several devices, several days, one conclusion.

That is also the polite thing to do as a finder. A log that says “found it, about eight metres north-east of the posted coordinates” is worth more to the next person than a smiley, and it gives the owner something actionable rather than a complaint.

Averaging, and why owners should do it

There is one habit that improves the coordinates on every cache an owner places, and most do not use it: averaging.

Rather than taking a single reading at the container, take several over a few minutes, or use the averaging function if the device has one, and publish the result. It costs a few minutes once and removes a share of the error that every future finder would otherwise inherit.

A cache with carefully averaged coordinates is quietly better than one without, and nobody who finds it will ever know why.

Why the arrow spins when you stand still

A compass arrow that rotates wildly the moment you stop walking is the single most reported symptom in this hobby, and it is two different mechanisms wearing the same costume.

The first is that most phones derive heading from movement rather than from a magnetic compass while you are walking. Standing still removes the input, so the arrow has nothing to work from and begins reflecting noise in your position instead. This is why the classic advice, walk ten metres past the point and turn around, genuinely works.

The second is magnetic interference. Phone compasses are easily disturbed by ironwork, reinforced concrete, vehicles and, routinely, by the magnetic clasp on the phone’s own case. A compass that reads confidently in the wrong direction is more misleading than one that visibly cannot decide.

Neither is a fault. The practical response is to stop treating the last ten metres as a navigation problem: note where the distance bottomed out, put the phone away, and search the ground, reading the ground is a skill the craft pages describe from the other side, as the hider’s work.

Reading the accuracy figure honestly

Most apps show an accuracy estimate, and it is worth knowing what kind of claim it is.

It is a statistical statement about your receiver’s confidence: not a guarantee, and not a statement about the cache. A reading of three metres does not mean the container is within three metres of the arrow. It means the phone believes it knows your own position to about that, and the listing’s coordinates carry an entirely separate error that no figure on your screen can see.

The two errors add. That is why a cache can sit reliably eight metres from where every finder’s device says zero, with nothing broken anywhere in the chain.

A third error is not an error at all but a transcription mistake, and it is far larger: reading decimal minutes as decimal degrees puts you tens of kilometres out. The coordinate converter explains the two checks that catch it.