# How to estimate the weight of the kilo?

Wikipedia has an article on the International Prototype of the Kilogram, which also includes a graph of the relative weight of its copies. Because there is (were) no independent unit to measure weight, every copy's weight were measured against the International Prototype.

It seem that most of the copies gained weight relative to the kilo over time, but it seems more likely that it lost weight¹. Considering the chart below, how would one make an estimate of the weight of the International Prototype in 1989 relative to its weight in 1889? Is it as simple as assuming a Wiener process and taking the average, or are there other things to consider? (The label numbers denote the other prototype copies, not weight.)

1: But it's also possible that it gained weight, just less so, or that it were in fact stable.

• W here is the chart from? – Aksakal Jan 11 at 22:47
• Why would a Wiener process be a reasonable model for weight changes? – whuber Jan 12 at 0:41
• @Aksakal From Wikipedia: en.wikipedia.org/wiki/… – Frank Vel Jan 12 at 10:30
• @whuber No particular reason, they could all have been gaining mass at different rates, and hypothetically have doubled in mass since 1889. But if they were drastically changing weight, I would have expected the weight dispersion in 1989 to be larger. So a Wiener process seems like a plausible, but very naive guess, as it disregards the trends of each line. – Frank Vel Jan 12 at 10:48