As if we didn’t need more drama, here’s an event that could prove to be pivotal: Philadelphia area transit has gone on strike and there’s no injunction. What I don’t for the life of me understand is how the bloody fuck SEPTA, presumably a public union, arrives at the idea that giving a Republican a chance to become president helps their cause. Seems like the absolute last thing they’d want.
I’m telling you now. Pennsylvania, even without this strike, is going to be closer than people assume.
A PAC has offered to fund free rides on Uber or Lyft to anyone who needs a ride to the poll, so maybe the impact won’t be so bad.
You know, we talk about all these events or tricks that could reduce turnout, but in reality none of this has as much effect as the weather. Snow or heavy rains in one place can wipe out all the human factors in several races, if not all of them put together. So what I want to know on Monday is, who is getting bad weather on Tuesday?
This weather-based predictive map that I found is quite informative (the map on the right side):
http://www.weather.gov/forecastmaps
It shows Arizona, Florida and the southern part of Texas going to Trump; Maine & surprisingly parts of Wyoming and Montana are going to Hillary. And if you think that’s a shocker, most of the rest of the country seems to be swinging toward Jill Stein and write-ins for the Liberal Democrats.
Ok, a serious answer on the weather.
Go here:
http://www.wpc.ncep.noaa.gov/basicwx/day0-7loop.html
You can stop the animation, and step forward to the frames (currently 5 & 6) that show Tuesday.
There are high pressure systems (fine weather) over the bulk of the east and west of the country, with rain ahead of a cold front that extends from TX northeast to the Great Lakes. But the only heavy rain & thunderstorms are over the solid red block TX, LA, OK, AR, KS, MO. Lighter rain over solid blue states further northeast later in the day. The only state that’s in play that will see any weather looks like IA, that will probably have a damp morning.
I was around when the Clintonistas boned them, but didn’t know they had actually been expelled.
So, where the weather is daunting and uncomfortable, the early voting is more likely to carry the day?
No, where the weather is daunting Democrats stay home, Republicans brave it.
That’s just hyperbole, but here’s an actual cite on how much weather can affect an election:
In general I would imagine weather reduces the correlation of the polling errors given the geographical diversity of the swing states. IOW the weather may hurt Trump in some states and Hillary in others; overall that may be good for Hillary given the current map though the effect is probably small.
This articlelooks at some of the research and the conclusion is that the weather probably won’t matter much.
“Deal reached to end [SEPTA] strike.” Wow, that was close. I really think that might have made the difference – just the sort of unpredictability that makes Nate Silver look wise.
Talk about leverage. I assume the strike organizers knew exactly how big a ticking bomb they had built, so to speak. Pretty brilliant, really.
FTR, I still think Hillary would squeak it out, strike or not, but it would definitely raise my blood pressure a little if they continued to strike tomorrow.
Wow, Clinton down to 63.6%.
But the betting market is still at 83%, and the rallies in the stock market and the $ have held up, so weight of money reflects a significant drop in Trump’s chances on the FBI announcement that has not shown up at all on 538.
The betting markets seem to think Clinton’s going to win decisively, with over 320 electoral votes.
Interestingly though, they seem to agree more with 538 on the Senate and House. 60% chance of Democratic takeover of the Senate, but the House is grim for Dems: the two most likely outcomes, totalling 98%, are Republicans 230-240 and GOP 240-250. The Republicans currently have 247 seats, so the betting markets are betting hard on not too significant losses there.
AS of 35 minutes prior to this it’s 65 and change.
Anyway, Silver’s model is wrong now, of that I’m certain.
Well, it’s not wrong. It works for what it does, it’s just a model on a computer. But it doesn’t incorporate early voting information, which at this point is very clearly favourable for Clinton. The state we know most about is Nevada, which 538 has as a toss up but is at worst 95% likely to go Clinton; that does change the forecast if you figure that in.
Yeah, 2016 has shown some flaws in the 538 model. Even Silver admitted on his podcast that it breaks down at the margins. In addition, as you noted, it does not take early voting into account, which is kind of odd for an aggregation model that is to forecast the election - why not incorporate the one “poll” that actually matters?
I understand it would be difficult to do as each of the States have differing requirements as to what they pre-announce, however, NC breaks down their voting numbers daily, which should be incorporated at least into the NC model.
Yes, this is really strange. Has he explained why?
I can understand that with the diversity of voting protocols there may note yet be a large enough sample size to develop an accurate model for the significance of early voting, but surely it’s obvious that it contributes significant information? Surely you must proceed either by incorporating the best provisional model you can come up with, or just stop forecasting at the time when early voting starts. It makes no sense to me that he continues “real-time” forecasting and just ignores it.
Early voting isn’t very predictive, maybe that’s why he doesn’t incorporate it.