Averaging the Way to Inadequate Information in a Varying World

At the recent Benguela 86 Symposium one of the participants decided to make a very strange recantation. There was sufficient evidence, in his view, to suggest that there was no reason to do the causal research in fisheries-related marine ecology, once the conventional average fishery information or parameter estimates were available. You could be right more often using average expectations in your data than if you used any three random variables with combined explanatory capabilities of up to 75%. He then proceeded to exemplify his conclusions from his analyses.