I should note that per the EY’s request I haven’t read the other comments before posting, so sorry if I duplicate anything.

The ability to make predictions in advance seems like one of the most important important, and assuming that you have enough time easiest to test measures of rationality. For the experimental and potentially the organizational level success on the prediction markets seems like an obvious choice, that also has the benefit of showing how good the person is at avoiding certain money related biases. There would of course need to be some controls in terms of equal access to capital and access to information, but I think we can work that out.
At the reputational level set up something like the David Brin prediction wiki for EY http://earthbydavidbrin.pbworks.com/w/page/15607657/Predictions though in this case we would be focusing on ones he explicitly makes rather than stuff culled from a work of fiction.