One could argue that the prospective approach is dumb, but for us this was important -- we wanted to know from the outset of the discovery process that replications would occur. This could lead to the findings being unusual in other ways...

For example, we might have picked trivial or easy to predict things to discover, thereby making it uninteresting to find that they are replicable. We did two studies with our findings to examine this possibility. In the first, we observed that researchers were not great at predicting the outcomes.

I really like this model. We would still need to incentivize the discovery process for tenure, grad students etc.