The brain trust over at FiveThirtyEight.com has held one of their Fivethirtyeight Chats and found that Donald Trump is not likely to shine in tonight’s debate. They write:
micah (Micah Cohen, politics editor): It’s been awhile since we’ve played “Buy/Sell/Hold” for the GOP primary. So, let’s play! The rules, as always: I give you the chances for each candidate as implied by the current betting line on Betfair.com, and you have to decide whether to buy/sell/hold stock in that candidate given that price.
OK, Nate and Harry, we’re going in reverse alphabetical order! First up: Donald Trump with an 18.7 percent chance to win the GOP nomination — buy/sell/hold?
natesilver (Nate Silver, editor in chief): First of all, I just want to say that as someone with a late-alphabet last name, I support reverse alphabetical order.
harry (Harry Enten, senior political writer): Very nice. And as someone with a brain, I support selling Trump at 18.7 percent.
natesilver: I’m a sell at 18.7 percent, indeed. Trump has risen quite a bit over at Betfair over the past month or so, which I’m not sure makes sense.
harry: It’s people seeing the calendar go by and Trump still doing decently well in the polls.
micah: Yeah, isn’t the logic something like, “The longer he sits atop the polls, the more REAL Trump’s support is”?
natesilver: But as we emphasize a lot around here, political time is not linear. What you say is true narrowly speaking. I don’t think anybody, us included, is saying Trump has literally zero chance. And those chances will rise the longer he maintains his position in the polls. But polls in early November aren’t particularly more informative than those in October or September or August or July.
harry: Most people are planning Thanksgiving and their Christmas shopping. They still aren’t paying all that much attention.
micah: I vaguely remember one of you saying in a previous chat to start paying more attention to the polls after Turkey Day?
natesilver: Yeah, we’ve mentioned Thanksgiving as a point when the polls start to get more interesting. Historically, that’s about when you begin to see much more widespread interest in the campaign.
harry: We still had time for a Newt Gingrich boom and bust after Thanksgiving last election, and the Iowa caucuses are a month later in 2016 than they were in 2012. Anyway, the point is that Trump is vastly overvalued at 18.7 percent for someone who has such a low net favorability ratingamong Republicans.
natesilver: re: Thanksgiving, I’m just saying that’s an inflection point at which the reliability of the polls begins to increase. Not saying that they instantly go from meaningless to meaningful. Here let me draw a graph.
[Nate gets up and draws on his office whiteboard.]