Why this measures a real sample instead of looking up a style
Searches for "how much does a gold chain weigh" mostly lead to tables: grams per inch for curb chain, rope chain, figaro chain, each broken down by a named gauge. Those tables look authoritative because they are specific, but the specificity is the problem — "curb chain" and "2 mm gauge" both cover a real range of actual wire diameters between manufacturers, so a table built from someone else's sample is, at best, in the right neighbourhood for a chain it has never seen.
A length actually measured and weighed has none of that uncertainty. If 5 cm of a chain weighs 1.2 grams, that is not an estimate of anything — it is a direct fact about that specific chain, and the only assumption needed to extend it to the full length is that the chain does not change gauge or pattern along the way, which is true for almost every plain chain sold by the metre or the inch.
The arithmetic, and when it holds
The method is linear extrapolation: divide the sample weight by the sample length to get grams per centimetre, then multiply by the total length. A 5 cm sample weighing 1.2 grams is 0.24 g/cm, so a 45 cm necklace of the same chain comes out to 0.24 × 45 = 10.8 grams. Nothing about the calculation depends on the chain's style, gauge or link pattern — it works identically for a fine box chain and a heavy Cuban link, because it never needs to know which one it is looking at.
The one thing that breaks the assumption is a chain that is not uniform: a clasp (usually a different, often base, metal), a soldered pendant bail, or graduated links that get thicker or thinner along the length. For a plain chain of one gauge throughout — most necklaces and bracelets sold by length — none of that applies, and the extrapolation is exact arithmetic on a real measurement rather than a guess.
A practical way to take the measurement
The cleanest sample is a spare length: a clasp repair offcut, a jump ring supplier's scrap, or a short length cut specifically to check. Lay it flat against a ruler rather than measuring it curled, since chain drapes shorter than it measures straight, and weigh it on a scale that reads to at least one decimal place — a kitchen scale reading whole grams only will not give a useful per-centimetre figure on a short sample.
Where there is no spare length, the same arithmetic works run the other way: weigh the entire finished chain, measure its entire length, and that becomes the "sample" for calculating the rate — useful for checking whether a chain matches its stated weight, or for working out the rate to compare against a different chain later.