Unfortunately, there just aren't many good metrics for future potential scientific value. It's not like you can just give NSA / DARPA / NIH one impact factor, as weird as that sounds.