Gear
Every instrument, and what each costs you
Rigid ruler, cloth tape, string, paper strip, calliper, photo, AR app: what each can measure, its typical error, and the failure it is prone to, in one table.
Instrument choice is not the largest source of error in a self-measurement, and it is worth saying that plainly before the rest of this post, because a comparison table can accidentally imply the opposite. Technique dominates. But instruments are not interchangeable, and each one has a specific, describable way it fails, which is worth knowing in one place rather than discovered by trial and error.
The reference table
| Instrument | Measures | Typical resolution | Bias direction | Prone failure |
|---|---|---|---|---|
| Rigid ruler | Length | ~0.5 mm | None inherent | Zero offset, angle, hair at the base |
| Cloth/fibreglass tape | Girth | ~1 mm | Reads high if loose or stretched | Sag, over-tension, stretch over time |
| Steel retractable tape | Neither, well | ~1 mm | Reads high (curved blade lifts) | Cannot wrap flat, hook slack |
| String + ruler | Girth | ~1 mm | Reads high if pulled | String stretch, mark width |
| Paper strip | Girth | ~1 mm | None inherent | Overlap mark precision |
| Callipers (single diameter) | Width, not girth | ~0.1 mm | Reads low as girth | Non-circular cross-section, jaw compression |
| Photograph | Neither directly | Not applicable | Unpredictable | No scale reference, perspective, lens distortion |
| AR measuring app | Rough length | ~1 cm on hard flat objects | Unpredictable, worse | Fails on soft, non-planar, non-rigid subjects |
Read "bias direction" as the way the instrument errs when used correctly, not as a mistake - some instruments have a built-in tendency to read one way, independent of the user's care.
Rigid ruler: the length standard
A rigid ruler is correct for length because length is bone-pressed - the ruler is pushed into the pubic bone until it stops, and only a rigid material transmits that pressure without bending. Its resolution is limited less by the millimetre markings than by how precisely a person can align an end to the bone and sight the tip through parallax-free viewing. The two failure modes worth knowing are a scale that does not start at the true end of the ruler, and hair at the base preventing full contact with bone. Choosing one that avoids both problems is a short checklist, not a purchase decision.
Cloth and fibreglass tape: the girth standard
A flexible tape is correct for girth because girth is a circumference and a rigid instrument cannot wrap one. Cloth tapes stretch with age and use; fibreglass-reinforced tapes barely stretch at all, which matters because a stretched tape has silently drifted from its printed scale and will read every girth slightly long from then on. The material differences between tape types are worth knowing before buying one, and any tape already owned is worth checking rather than trusted.
Steel retractable tape: the wrong tool for both jobs
A retractable steel tape resists curling into a small circle, so it sits away from the surface along the wrap and reads the girth high, and its rigid, curved blade cannot press to bone for a length reading either. It is a case where an instrument that is excellent for its intended purpose - measuring a room - is simply the wrong shape for this one, not a lower-quality version of the right tool.
String and a ruler
Wrapping string around the shaft, marking the overlap, then laying the string against a ruler works in principle and is a legitimate fallback. Its two failure modes are string stretch under the tension used to wrap it, which is not controllable in the way a tape's tension is, and the width of the mark itself, which is a source of error on a thin string in a way it is not on a printed tape. The method and its two specific errors are worth reading before relying on it for anything you intend to compare to a published figure.
Paper strip
A strip of ordinary paper, wrapped and marked at the overlap, then read against a ruler, reaches essentially the same accuracy as a tailor's tape for girth. It has a genuine advantage over a tape: wrapping and reading are separated into two steps, so there is no moment where a hand is simultaneously holding tension and trying to read a number, which is where a meaningful share of tape-reading error actually happens. Its only real failure mode is imprecision in marking exactly where the overlap sits, which a fine pen and a careful hand largely solve.
Callipers: the wrong quantity entirely
A calliper measures a diameter - the straight-line width across an object - not a circumference. Girth is a circumference, and converting a single diameter reading to circumference by multiplying by pi assumes a perfectly circular cross-section, which a shaft is not; it is closer to an ellipse, and a single-diameter conversion under-reads the true circumference as a result. The geometry of why a circle-based conversion misses on a non-circular shape explains the direction and rough scale of the error, and a calliper jaw also compresses soft tissue as it closes, adding a second source of downward bias with no natural stop to prevent it. A calliper is a legitimate instrument for a genuinely different question - the width of something rigid - and not a substitute for a tape here.
Photographs: no scale information
A photograph, on its own, contains no length information at all. Anything derived from one - including a ruler visible in the frame - is subject to perspective, lens distortion, and the angle of the reference object relative to the subject, all of which change the apparent scale in ways that are difficult to correct for without knowing the camera's exact geometry. The specific ways a photograph corrupts a length reading, even with a ruler in frame, are worth understanding, because "I had a ruler in the picture" feels like it should solve the problem and does not fully solve it. This is also the honest answer to why an image-based service cannot output a length in centimetres at all - it can output a rating, which is what a platform like Rate Cock is actually built to do, scored the way a dedicated photo-rating tool scores an image, and that is a different and legitimate kind of output rather than a lesser attempt at this one.
AR measuring apps
Phone AR tools that estimate a length by tracking the camera through space resolve to roughly a centimetre under favourable conditions: a hard, well-lit, planar, rigid object held steady. None of those conditions hold for this measurement - the subject is soft, not planar, and cannot be held rigidly still relative to a slowly moving phone - so the resolution in practice is considerably worse than the marketing figure for hard objects, and the failure mode is unpredictable rather than a consistent bias in one direction. What AR apps can and cannot resolve, and why goes into the tracking method itself.
What "resolution" does and does not tell you
The resolution column in the table above describes the finest distinction an instrument can register under ideal conditions, and it is easy to over-read. A digital calliper resolving to a hundredth of a millimetre is not a hundred times more useful than a ruler resolving to half a millimetre for this particular measurement, because the thing being measured compresses under its own handling by an amount larger than that difference. Resolution only helps up to the point where something else in the chain - technique, tissue compression, state - starts contributing more noise than the instrument's own scale does, and for every row in the table above except the calliper, that ceiling sits well above the instrument's stated resolution already. The calliper is the one entry where the instrument's precision genuinely exceeds what the measurement itself can support, which is a caution about the instrument rather than a point in its favour.
Checking whatever gets chosen
None of the resolution or bias figures in the table above are fixed properties that hold forever once an instrument leaves the shop. A tape stretches with use, a ruler's end wears down, a printed scale can be reproduced slightly off true by a printer that does not default to actual size. Checking an instrument against a known reference - a standard payment card is a convenient one, at a fixed 85.60 by 53.98 millimetres - takes under a minute and catches a drifted scale before it quietly becomes a wrong number recorded as fact. The complete checking routine, covering every instrument type on this page, is worth doing once for any instrument that will be used more than a handful of times, rather than assumed accurate indefinitely because it was accurate on the day it was bought.
Reading the table as a whole
The pattern across every row is the same: an instrument built to wrap reads girth, an instrument built to stay rigid reads length, and every attempt to make one instrument do the other's job produces a specific, describable, non-random error rather than a smaller version of the right answer. None of the entries above claw back much accuracy over a plain rigid ruler and a plain cloth tape - the technique used with either one matters far more than which specific model of either is in your hand, and where the largest share of self-measurement error actually comes from is worth reading as the companion piece to this table rather than assuming the instrument was the weak link.
A separate category of tool is not attempting to measure at all, and should not be judged against this table. An AI system estimating from a photograph, explained in full elsewhere, is doing pattern recognition rather than measurement, and a human judge's impression, described on its own terms, is doing something else again - both are legitimate for what they are, and neither belongs in a table about centimetres and their error bars.