The best way I have seen this applied was at a call centre (had to right the scripts for it).

It worked better because it was a giant sample size of employees (making it more likely that their abilities were normally distributed), their performance did not depend on one another, and evaluation was done with actual metrics.

But I have seen similar theory applied to 'teams' as small as 2, with zero metrics beyond manager opinion. And when it is used like that it is basically a straight-jacket that is both unfair and has unintended consequences.

And what it truly reflects is leadership not trusting their managers to make honest assessments of performance.