It's remarkable that there is data about the outcomes of billions of lives, but we still have only vague indications about which behaviors lead to which outcomes.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3730121/
> We demonstrated that children’s sustained decisions to wait for a greater reward rather than quickly taking a lesser reward are strongly influenced by the reliability of the environment (in this case, the reliability of the researcher’s verbal assurances). More broadly, we have shown that young children’s performance on delay-of-gratification tasks can be strongly influenced by an implicit rational decision-making process.
Do you really find that remarkable or is that just hyperbole? The fact that its so difficult to do controlled experiments - the holy grail of science - in this area in tandem with the fact that these are almost invariably complex systems we're dealing with makes it such that its not that surprising to me that we are still as confused as we are.