Interesting to see what happens with Sam Wang's Princeton Election Consortium at the next update

[Loki Liesmith]

UPI/CVoter (Moonies and some Brand-X Indian public opinion outfit) just dumped a bunch of shitpolls for all 50 states.

It will be interesting to see what this does to Wang's numbers as Pennsylvania is heavily leveraged (a swath of recent, suspect close or Trump leaning polls) in the last week or so.

The one thing I like about Wang is that his methodology resists outliers. But sometimes you get a sequence of bad polls. It happens. It knocked my model off a few points. I'm hopeful his is more resistant.

(on edit: no I don't think it's real. It's just an interesting opportunity to test a hypothesis).

[Farmgirl1961]

1. Is that why Nate Silver 538 model actually dipped a little?

n/t

[Loki Liesmith]

2. Those polls aren't in the top 10 most weighted polls in his update

so doubt it.

It's a C rated poll.

But frankly, I refuse to work with 538 numbers any more. See here.

[Loki Liesmith]

3. Only knocked it down 2 points

PA is more heavily leveraged in my model. Maybe want to adjust.