The probabilistic forecasts debate -- That's nice fellows, but I'm kinda focused on that iceberg

Andrew has a good breakdown of the debate over probabilistic election forecasts if you're interested. Here's why I think you shouldn't be.

1. At this point, economists have so thoroughly screwed up the discussion of rationality that I now make a conscious effort to avoid the term “rational,” much as I try not to use “significance” except in the technical sense. Rational actor arguments routinely rely on simplifying assumptions that are never observed in real life and which exclude most of what we would think of logical thinking.

2. That said, any reasonable “rational choice” model has got to address the accuracy and completeness of the information available to the actors. For people who relied on 538, perceived and actual closeness matched up fairly well and it makes a fair amount of sense to talk about “rational” voting behavior. For people who relied on NYT’s Upshot and Slate’s real-time exit polls, not so much…

3. With this in mind, this whole debate over probabilistic forecasts strikes me as another one of those secondhand smoke on the Titanic questions that are eating away at our collective decision making ability. We have larger and more immediate concerns.

Even with good data and solid analysis, it seems obvious to me that the social good of horse race coverage is U-shaped. A reasonable amount provides useful information and helps voters converge on a generally satisfactory choice. Too much (either probabilistic or old school) overwhelms the process and creates a noisy feedback loop. Even if we were getting good information we’d still be on the wrong side of the U, and we aren’t getting good information.

4. It’s convenient to blame the voters for being “irrational” but given the crap they’ve been fed, it’s difficult to see how they can reasonably be expected to make intelligent decisions. We have supposedly serious organizations like the New York Times giving oxygen to joke candidates.

Billionaire vanity campaigns crowding out viable options. Millions of early votes wasted because someone thought it would be a good idea to have Super Tuesday three goddamn days after the first real primary of the election (no one takes NH seriously). And, of course, a level of noise that would strain even a well running system.