Thank you Nevada, Iowa, Ohio, Florida and North Carolina - your support is greatly appreciated.
Posted on 09/26/2016 6:21:25 AM PDT by Ravi
Nate is coming around.
(Excerpt) Read more at projects.fivethirtyeight.com ...
I’ll hand it to Nate on one thing. He’s sticking to his methodology, even if he doesn’t like the results.
He is at about 55% in the now cast. He still hasn’t flipped the final state to be winning on the “snake” (states lined up graphically in order of probability) but everything on his side of the line in the now cast is at 67% plus, and Hillary has four (Colorado, Pennsylvania, New Hampshire and Michigan) that she has only a week hold on that she has to sweep—with the odds being that she won’t do that the majority of the time, and won’t pull in anything else to compensate.
His primary concern with his results is that they are accurate.
actually, Silver’s Nowcast projetion has Trump winning 55-45, with 276 EV’s; yet the coloration map he provides shows Clinton with 273...something wacky in his methodology...
See my post 3. Basically, if there are a bunch of states that statistically one is barely favoured in, one will loose one or two of them on average.
I don't have that luxury in my forecasting model. Based on the polls as they are reported, I showed Trump at 43% as of Saturday. Late Sunday polls have pushed that number up slightly to 44%.
We will see after the debate whether or not Silver's poll adjustment forecasts come true or not. For me, I expect some will but I choose not to get ahead of the data. I prefer to report what the data actually is, and let the reader draw conclusions on whethe the model is wrong on the data is rigged.
-PJ
Silver has funding, staff, and about a decade of experience tweaking his model. I’m sure that the number adjustment is based on a formula based on a number of factors (some regression to mean? Previous track record?) rather than gut. It isn’t just gravitas, though the gravitas allows him to use his way of tweaking without getting flack.
Exactly, he wants an accurate outcome. He’s not trying to affect the outcome by playing with his data.
So Silver is applying his "house effect" to the polls, essentially saying that the major pollsters are lying to the public.
To believe Silver, one must publicly call out the polling industry as dishonest, or he would not need to flip Clinton poll wins into Trump poll wins.
-PJ
There is a difference between being deliberately being dishonest and handling data sub-optimally due to incompetence. Sometimes the lines between these two are not clear. Silver spends a fair bit of time analyzing the problems with polling.
one will loose one or two of them on average.
so he’s essentially saying his own analysis (as regards pct of victory) is nothing sharper than any armchair analyst could do, after drinking a six pack of brew in twenty minutes...
Thank you Nevada, Iowa, Ohio, Florida and North Carolina - your support is greatly appreciated.
But as I said in post 11, he's effectively calling the other polls politically manipulated if he feels he has to flip most of the Florida polls from Clinton to Trump. It's not just one poll here and there, it's most of the Florida polls by different firms.
What does one make of that? How does one accept Silver's adjustments without publicly calling out the polling company to respond and defend their original numbers?
I'm not hearing anyone dispute Silver's adjustments. I'd expect some of the more reputable polling companies to complain and push back to defend their own accuracy.
-PJ
Let’s not bee to hasty about NC. Chelsea Clniton’s been making many “campaign stops” there lately, stumping for her mom, drawing “crowds” of 20-50 every place she goes... *snort*
I was polled two weeks ago. The pollster had a foreign accent, and judging by how many times we each had to repeat ourselves, I have no doubt he was calling from India, with a poor connection. I hung in there and answered all his questions...but odds are many don’t. The whole affair made me lose any confidence I had in the polling process.
I think the only thing that can be discerned is trends. If candidate ‘A’ keeps going up...he’s in good shape, and vice versa. But if its at all close, the actual percentages probably don’t mean much.
In the meantime, he'll do his best to preserve the "horse race" as long as possible.
I just see it as an opportunity to shame the traditional polling companies to stop gaslighting their polls. Maybe it will take a cage match with Silver to do it.
It makes me wonder if juicing the polls is part of the MSM’s “Trump is too dangerous for journalistic ethics” plan to do anything to win.
-PJ
Looks like the only change from 2004 is Virginia. If Trump can somehow get NM to his side, it’s over.
Disclaimer: Opinions posted on Free Republic are those of the individual posters and do not necessarily represent the opinion of Free Republic or its management. All materials posted herein are protected by copyright law and the exemption for fair use of copyrighted works.