The ideal I want to attempt to achieve for the purposes of this conversation is a resource allocation process that is based on the same scientific approaches as the projects themselves.

So, if by politics you mean "the act of people mutually deciding what to do with limited resources," then yes, you cannot eliminate politics from scientific funding. But that feels like a tautology to me.

If we use the colloquial definition of politics, namely "nepotism, influence peddling, backbiting over unrelated issues, earmarking, pork barreling, gerrymandering, etc." then I do believe we can and should strive to eliminate politics from scientific funding decisions.

Thanks for your detailed reply; I hope this clarifies each of our intended meanings.