New and improved, with even less ideology!
Based on various people's input, I've revised the system. The model now distinguishes between short-run impacts and long-run impacts, with different respective methodologies and weights. That provides more accurate scoring and also makes it easier to narrow it down to an evaluation of short-term EA causes (extreme poverty and animal welfare). The report now includes more evidence and issues, to the point where I judge that creating useful recommendations quickly is now more important than making further refinements. (The methodology / scoring cycle can continue with more revisions afterward, of course.)
So I am planning to move into actually evaluating candidates soon. I would like (a) further review of this document to suggest final changes and different estimates than mine, and (b) people who are willing to work with me on judging political candidates.
Also: I don't know what to call this project and I am taking name suggestions.