Science is a very stochastic process though. Yes, not a lot of research is eventually useful but that's because chance can play a huge role. If the success rate is 1%, then investing less will not necessarily increase the success rate. It will decrease the absolute rate of discovery.