All statistics have some context they were created for, most get used outside that context.
Unemployment rate is job-seekers/(job-seekers + employed). Long term job seekers should technically be included (many are) but when they can't find a job long term, they tend to re-classify as students, stay-at-home parents, early retirees, disabled...
In practice, that statistic is usually most relevant for estimating the number of people receiving unemployment benefit. The data collection is intimately tied to this: you don't need to tell anyone you're unemployed unless you are applying for a benefit. Places (eg Ireland) with generous unemployment packages will generally register a higher unemployemnet rate because people have more of a good reason to register as unemployed, even if they expect to be re-employed quickly. Places (eg US) with much more generous disability benefits than unemployment payments will have a lower unemployement rate as people (that are borderline cases) have a strong need/incentive to be classed disabled. Australia has generous student welfare programs, this reduces measured youth unemployment. Etc. ad infinitum
Basically, unemployment is a measure of (more or less) unemployment welfare recipients/applicants. It can't be used to compare between regimes.
It also happens to be an unreliable lead indicator for short term labour market changes. I think this is what has given it its current usage.
What the press should be reporting on is employment rate: Employees/18-65_Adults. This is not gameable, and it means the same thing in every time and place. It measures what people care about when they report on unemployment rate. Alternatively, they could also report "total salary" for the country. That's how most markets get measured (eg total oil sold, total market returns, etc.).