"Primary Colors": What the 12-Point Platform would look like if the political class had written it
Now, holding the Democrats accountable and dragging them kicking and screaming, whining and whimpering, to the left -- which is not necessarily the same as dragging the Overton Window itself left -- is not the worst idea in the world; and Atrios is, after all, my blogfather. So I took a look.
Since I wanted to find out who ran and funded the site, I looked at the About page, which is blank as of this writing. Then I checked out the FAQ:
Ryan and Jon [Geeting]! Ryan is a former Democratic campaign-staffer (who wanted to utilize his mathematics degree) and Jon’s a blogger who currently also runs the Pennsylvania liberal blog Keystone Politics. We’re also both PA-born, so don’t be that surprised if we accidentally focus on Pennsylvania often. Primary Colors was self-funded by the two of us and has no affiliation with any PAC, campaign, or organization.
(Jon Geeting is also the author of the Eschaton post linked to above.) So, far be it from me to load up the snark gun for a self-funded effort, especially one hailing from my old-time stomping grounds in Pennsylvania. I mean, I wouldn't want to take down the next Nate Silver with a careless shot. Kidding!
But I do have some concerns about their methodology. Look again at the screenshot of "Good Democrats" above. Each Democrat has an Expected Score, an Actual Score, and a Progressive Value, which (positive or negative) is the difference between the two. For example, Raul Grijalva's Expected Score is 84.2, his Actual Score is 97.3, and so his Progressive Value is +13.1. In other words, he's more progressive than you'd think, from his District.
But how are the scores derived? Again, the FAQ:
Our scores come from a variety of sources. They are a composite of a few different rankings of members of Congress, and we give a heavier weight to Progressive Punch (PP) scores, since we feel they best personify how progressive each member of congress is. We also average in DW-NOMINATE scores and Party-Line votes, but those more measure levels of ideology and partisanship, respectively — but we feel the PP scores do a better job of capturing what is or is not a progressive vote. While we see value in partisanship, as ideological liberals we’re more interested in grading members on their actual issues positions. We also used the Cook Partisan Voting Index (CPVI) and our actual score (PPCI) to determine where members of congress should vote based on their colleagues scores in similar districts, which we call their “Expected Score.” Subtract the “Expected Score” from the actual score, you’ll get a member’s “Progressive Value” — the measure of how beneficial a member is to the progressive movement based on their state/district lean. Finally, based on the Progressive Value, we assign each member a Primary Score from 0-10, with zero meaning “Don’t Primary” and 10 meaning “Must Primary,” giving a little extra slack to those members in R+ districts.
(There's also a more extensive explanation at Methodology.) In other words, the Primary Colors Score is an indexing strategy: It's meta: an index of indexes that synthesizes three other indexes, enhanced with additional data. (Though I don't want to take the financial metaphor too far, one can imagine "passive investors" saying "Put my contributions only into campaigns with a Progressive Value of +4 or greater." No more rubber chicken dinners!) And a synthetic index is only as good as the indexes it aggregates, and each of those indexes is only as good as the data it's a proxy for. So, let's look at Progressive Punch's indexing methodology for, oh, health care, first. Here it is:
See "single payer" on there anywhere? No? So what kind of "progressive" "punch" are we talking about here? Not Frazier's or Ali's, that's for sure.
Take their second index, DW-NOMINATE. Since I'm not a political scientist, I'll have to rely on WikiPedia:
Though there are important technical differences between these types of NOMINATE scaling procedures; all operate under the same fundamental assumptions. First, that alternative choices can be projected on a basic, low-dimensional (often two-dimensional) Euclidian space. Second, within that space, individuals have utility functions which are bell-shaped (normally distributed), and maximized at their ideal point. Because individuals also have symmetric, single-peaked utility functions which center on their ideal point, ideal points represent individuals' most preferred outcomes. That is, individuals most desire outcomes closest their ideal point, and will choose/vote probabilistically for the closest outcome.
Again, I'm probably opening myself to being beaten like a gong by an actual political scientist, but at least as far as policy goes, doesn't a two-dimensional Euclidean space of "choice" preclude, a priori, the very concept of "strange bedfellows"? Ron Wyden (D) and Rand Paul (R) can agree on Constitutional issues, after all. And I've seen Republicans (though not elected ones) make the case for single payer. For example, with the 12-Point Platform:
8. Enforce the Bill of Rights
9. End the Wars
very clearly appeal to at least some factions on the "right" as much as those on the "left." So, again, from the standpoint of progressive policy ("progressive" being an ill-defined term at best, even assuming gopd faith) it's not clear how adequate a proxy for progressivism DW-NOMINATE really is.
Finally, on the Cook Partisan Index:
The Cook Partisan Voting Index (Cook PVI) is a measurement of how strongly a United States congressional district leans toward the Democratic or Republican Party, compared to the nation as a whole.
Which assumes, of course, that being a Democrat is an adequate proxy for being a progressive. Could be, but I wonder how many people (especially influencers) have left the Democrats in disgust or despair, or were thrown out, as in 2008.
So, an index of indexes is only as good as the indexes that it synthesizes, and they, in turn, are adequate proxies for what they purport to represent insofar as their components allow them to be so. (That is, the Dow Jones average would not be an adequate proxy for buying or selling stocks if it included buggy whip manufacturers as a component. Similarly, if the health care component of PP alone was regarded as an adequate proxy for the presence of progressives in Congress, and it included a heavily weighted single payer component, the PP would be very low; perhaps as low as zero. Some would say that's accurate!) I don't question PC's knowledge or professionalism, and I'm sure these are the best indexes available. But just because they're available doesn't make them fit for their purpose! 
Finally, however well-conceived the "Primary Colors" indexing strategy might be, I don't think it gets at the central problem of politics in America today: Concrete material benefits for the voter. Once again the FAQ:
Absolutely not. We partially understand the comparison, though: If you look at what we’re doing at a glance, it seems like we’re giving Democrats a litmus test and threatening to primary anybody in the Democratic party that is not up to snuff from the left. Though the people who think that this is true clearly haven’t read much about our site or objective. Our scores are based on an algorithm that measures how members of congress should be voting based largely on their colleagues in similar districts, then assigning them a ‘primary score’ based on how far above or below they are their expected score. Unlike the Tea Party primaries, this isn’t a personal ideological agenda from us against everybody with a primary score. We wanted to give progressives a very clear sense of how their MOC is voting, along with how you should expect them to vote based on their district’s partisan lean.
But litmus tests are exactly what we want! "Partisan lean" is for the rotisserie Inside Baseball types who front-page at Kos.
Primary Colors is all about the meta. And the meta is the wrong metric. There is only thing that matters: policy. Only policy brings concrete material benefits for voters, along with any structural reforms needed to bring those benefits into being. And a series of litmus tests -- a checklist -- like the 12 Point Platform provides a simple and proven metric to hold politicians accountable for policy. "More and better Democrats" doesn't mean squat if nobody knows what "better" means!
So, you can have a simple and transparent 12-Point checklist, designed to change what's "politically feasible," or you can have an index of indexes standing between you and the results you really want, with the political class running the show based on an opaque algorithm, and all designed to accept what's "politically feasible."
Which'd you rather?
NOTE  I assume the reference to Joke Line's roman a clef about the 1992 Clinton campaign is accidental?
NOTE  Please don't hate me because the execrable Chris Bowers also hails from Philly.
NOTE  What progressive "movement"? Motion, a la a quivering blanc mange or perhaps a nice custard (or a trifle), doth not a movement make!
NOTE  The Dow Jones Industrial average is a putative index that serves as a proxy for health of the economy or at least the stock market. Up is good; down is bad. The DJIA is a weighed average of the stock prices of 30 companies ("components" of the average). Every so often, the components are changed so that the index retain its value as a proxy. U.S. Steel was removed from the Dow Jones Industrial Average in 1991, for example, because nobody cares if we make steel, any more than anybody cares if we make buggy whips.
NOTE  See the Streetlight Effect.
NOTE One imagines how the Progressive Punch index would treat slavery: There would be items for the weight of shackles, the distance a slave could be forced to walk per day in a coffle, once a year's change of clothing, etc. And then "progressives" would be ranked on whether they supported light-weight shackles, or heavier ones, or were against regulation entirely. Oh, you think you're not at that point?