Robert likes this. Well, at least the first 8 pages, before the math starts.

The Limits of Meritocracy
John Morgan, Justin Tumlinsony. Felix Várdyz
Abstract
We show that too much meritocracy, modeled as accuracy of performance ranking
in contests, can be a bad thing: in contests with homogeneous agents, it reduces
output and is Pareto inefficient. In contests with sufficiently heterogeneous agents,
discouragement and complacency effects further reduce the benefits of meritocracy.
Perfect meritocracy may be optimal only for intermediate levels of heterogeneity.

Downloadable here.

Robert thinks of his experience with high school swim team.  Very unenjoyable because it was clear that of the 8 standing on the starting blocks, one was almost certainly going to win.  Maybe a second guy had a small chance.  But nobody in the other 6 blocks had a chance.  So why try?