If you base it on trying to measure biological characteristics that we know (or think we know) affect performance, you are right--it would be impractical for all but leagues that the top athletes compete in.
If you made it based on past performance, though, it should be doable in even small amateur leagues.
That's how it works in chess. When you play tournament chess you get a rating based on your performances and the ratings of the established players you played.
Large tournaments use the "Swiss system". In round 1, you sort the players by rating, split the list in two, and the top rated player in the first list plays the top rated list in the second, and so on.
In subsequent rounds, you sort the players by results so far, split the list into separate lists for each score group, and then within each score group the top half by rating plays the bottom half.
Large tournaments typically have prizes for the top scores, but also "class" prizes. A class prize is for the top scores in certain rating ranges, say under 1800 or under 1200. A big tournament like the US Open might have a whole slew of class prizes, such as top score in under 1200, [1200, 1400), [1400, 1600), [1600, 1800), [1800, 2000).
There are women's chess tournaments but that is because much fewer women take up chess, and often women are uncomfortable if they are the only woman in a competition, especially at the scholastic level.
That environment can affect their performance due to "stereotype threat". When you are doing something in public that goes against a negative stereotype about your group you can feel a lot of pressure to avoid confirming the stereotype, and that extra pressure can actually make it harder for you to do your best. That in turn further discourages women from taking up tournament chess.