How scientific impact scales with funding

So yet another grant proposal bit the dust this week, just an eye roll rather than total devastation, the show must go on. However, I wanted to give this little beauty of a paper some extra air time. It tells us what we may suspect already, that giving the same groups more and more money does not necessarily increase the likelihood of their getting more and more good data. The authors conclude that funding strategies that target diversity, rather than 'excellence' are likely to be more productive.
For me it seems like there are excellent groups doing cutting edge research in their special areas of interest with millions in grant funding, yet we are still seeing very little in the way of therapy for the diseases they work on. It must be time to make a decision to fund small groups with new ideas, and for more than just a year at a time.
To say that an organisation 'funds the best science' suggests we can see into the future...sometimes it's the quirky off-beat ideas that turn out to be the best.