The reward system also prevents dead ends from being identified, publication of approaches that did not lead to the expected results or got nul results, publishing confirmations of prior papers, etc.

Basically, the reward system is designed to be easy to measure and administer, but is not actually useful in any way to the advancement of science.