patrick.net

joshuarito is really jizzed up over the outcome of a coin toss.

My take on Nate Silvers work this election:
1. He gave Trump a much much higher chance of winning than the Newspapers, which used models with bigger assumptions. So he gets credit for that.
2. Based a lot on reading Silver's blog I said that Hillary would probably win the popular vote by 1 to 7 pts. She won by just outside of that (0-1 pts), so I was mildly surprised, but not hugely The reason that I picked +-3 points as reasonable is that polls often miss by 2 pts. I gave an extra point swing due to the 'different' nature of this election, the challenge pollsters had in figuring out likely voters in this election, and the fact that people filter their calls more due to spam calls.
3. Silver gave Trump a 30% chance of winning the electoral college. That is not too far off from even chances - it's two coin tosses out of 10 away from 50/50.
4. Lots of idiots will get the result of a coin toss correct. That doesn't make them a genious. Getting many right in a row is necessary for that. Ironman is now 1 for 1 on his coin-toss predictions while I've been watching.

Ironman's main argument was that the polls oversampled dems by about huge amounts (often > 10 pts). Yet, the result was off by only 3 pts. This is pretty consistent with past polls and election results. The 2012 results were also off by 3 pts, but in the other direction.

When you go state by state and look at how many states were toss ups, you see that a 1 or 2 pts swing can make a huge difference in the electoral college. When person who the polls predict to win outperforms the polls, then the state by state predictions end up good (2012). If the person who the polls say will win ends up under-performing (2016) then the state by state results look terribly wrong. That fact is very predictable.

People were surprised by the results, but they shouldn't have been shocked, especially if they had been reading Nate's blog at all.

That said - nice work Ironman. I don't agree on the media skewing the polls conspiracy theory, but you got the results right, so you deserve credit for that.

YesYNot says

joshuarito is really jizzed up over the outcome of a coin toss.

No, it's just fun trolling you.

But your boy Nate was wrong (bigtime) :)

YesYNot says

3. Silver gave Trump a 30% chance of winning the electoral college. That is not too far off from even chances - it's two coin tosses out of 10 away from 50/50.

Omfg!

joshuatrio says

Ladies and gentlemen, I present to you, Nate Silver's biggest supporter.

Not being retarded hardly makes me a Nate supporter. You havent challenged anything I've said with anything but emotional nonsense.

You don't even comprehend what I said.

marcus is NOT deplorable says

It's also true that someone can make the most accurate possible assessments of probability, and guess what ? One third of the time they assess that something has a one in three chance of happening, it happens.

I'm not sure what's so difficult for you guys to understand about this. What ? You think with the state polls Nate he had access to, the same ones everyone else had access to, he should have been able to say that Trump was the favorite ?

marcus is NOT deplorable says

You don't even comprehend what I said.

Please take a moment to read a few snippets from our previous dialogue.

joshuatrio says

marcus is NOT deplorable says

Nate had Trump winning most of the primaries. His model gave Trump a better chance than many other analysts. He talked about the large numbers of undecideds and Hillary's electoral college problems. That's why he gave TRump such a high chance.

His credibility is in tact.If he were a pollster, then you might have a point.

Idiocracy is here folks.

Marcus, he was wrong. His models failed him. He went on his own bias and lost big (3 times). I'm sorry you butthurt liberals can't see through the fact that your "prince nate" isn't all he's cracked up to be.

joshuatrio says

Wow.... getting desperate eh? This WAS about guessing right. Otherwise, Nate wouldn't be in business in the first place.

I agree, started backpedaling near the end, but it was too late. Isn't Nate's claim to fame predicting every state for Obama? If so, he came no where close for this one. Hell, my prediction was closer than his was...

Let's see... he missed, Michigan, Wisconsin, Florida, North Carolina and Pennsylvania

I missed... Wisconsin. Holy shit! I should start a blog now.

One of the comments on poor Nate's site:

"I'm sorry but I just can't help feeling like you somehow let me down Nate. I'm still a fan, I still believe you're the best out there, but unfortunately in your business, you don't get to be so wrong and retain credibility...."

Yes, and CIC is a genius, becasue what he wished to be true was closer to what happened than what was concluded by good analysis.

joshuatrio says

marcus is NOT deplorable says

Nate had Trump winning most of the primaries. His model gave Trump a better chance than many other analysts. He talked about the large numbers of undecideds and Hillary's electoral college problems. That's why he gave TRump such a high chance.

Dumbass, I'd encourage you to go to the first page of this thread.

I understand your point, have already challenged it

I guess if by challenge you mean showing clearly that you don't comprehend what I'm saying.

marcus is NOT deplorable says

I guess if by challenge you mean showing clearly that you don't comprehend what I'm saying.

joshuatrio says

Marcus, he was wrong. His models failed him. He went on his own bias and lost big (3 times). I'm sorry you butthurt liberals can't see through the fact that your "prince nate" isn't all he's cracked up to be.

.......

Wow.... getting desperate eh? This WAS about guessing right. Otherwise, Nate wouldn't be in business in the first place.

I agree, started backpedaling near the end, but it was too late. Isn't Nate's claim to fame predicting every state for Obama? If so, he came no where close for this one. Hell, my prediction was closer than his was...

Let's see... he missed, Michigan, Wisconsin, Florida, North Carolina and Pennsylvania

I missed... Wisconsin. Holy shit! I should start a blog now.

joshuatrio says

No, it's just fun trolling you.

At least you realize that you are not saying anything useful, and won't stand by your comments. Stand back folks. Don't get jizzed on.

YesYNot says

At least you realize that you are not saying anything useful, and won't stand by your comments.

You fail at basic reading comprehension... And I've stood by my comments from the beginning... But yes, it is fun trolling you...

http://dailycaller.com/2016/05/04/7-times-nate-silver-was-hilariously-wrong-about-donald-trump/

http://www.thewrap.com/nate-silver-blew-it-bigly-on-the-election-can-his-brand-recover/

“Nate was arrogant. His numbers were all over the place. The title of ‘guru’ is now gone,” The Hill media reporter Joe Concha told TheWrap. He said Silver’s career will survive, but “never again will he be held in any revered regard.”

I'm sorry man, he was wrong on Trump. Quit being so whiny and accept it. He gave Clinton over 70% chance of winning and Trump walked away with over 300 electoral votes.

indigenous says

So then the bookies missed that the polls were rigged too? I mean if you can't trust the bookies who can you trust?

You really don't understand how markets work, in the least.

Bookies book wagers

Oddsmakers set opening lines

Punters move the market price

Bookies profit from vigorish

What bookies want is 50/50 split

Allow me to illustrate

Book offers

Who will win the 2016 presidential election?

Hillary -300
Trump +250

Now imagine they book $1 million on each side, and Hillary loses

Book banks $3 million from the hillary losses, and pays out $2,5 million to winning trump bets

They profit $500k

joshuatrio says

I'm sorry man, he was wrong on Trump. Quit being so whiny and accept it. He gave Clinton over 70% chance of winning and Trump walked away with over 300 electoral votes.

First off, I never held Nate in "guru" regard. But I think he factors in more details than anyone else.

For most of the 5 days prior to the vote, he had CLinton in the 64 to 66% range. I guess he ticked it up on Tuesdays morning.

But what you apparently will never understand is that given all of the information, that may have been a perfect analysis. You haven't said the slightest thing to indicate you even understand why this is true.

EVENTS WITH A 30 to 36% CHANCE OF HAPPENING , CAN HAPPEN !!"

THink of it this way, if there are infinite parallel universes, and NAte's analysis was perfect, and all the universes split of from being identical on Tuesday, then in about two thirds of the universes, HIllary won, and in the other third Trump won.

I'm sorry man, he was wrong on Trump. Quit being so whiny and accept it. He gave Clinton over 70% chance of winning and Trump walked away with over 300 electoral votes.

--------------

Math is not your strong suit

I was just trying to help you understand what probability means, but I know my multiverse example isn't going to help you, becasue obviously what's going to happen is baked in Tuesday morning.

There were things that can't be known, at least not from existing polls.

I think the problem is you never got what Nate does.

I never thought of Nate as a guru. And I always took his numbers as reasonable assessments of probability. This is why I was disappointed, but not shocked that Trump won. I know what a 35% chance of of an event happening means.

marcus is NOT deplorable says

But what you apparently will never understand is that given all of the information, that may have been a perfect analysis. You haven't said the slightest thing to indicate you even understand why this is true.

Brah, I understand that he gave Trump a chance.... like 29% (ish). Which is not even *GASP* 1 out of every 3 coin tosses!

YOU SEEM TO BE BLINDED by the fact that... he more or less predicted a Clinton landslide, not just with his model, but his own personal prediction, and has been laughing at Trump since the primaries.

Your boy "blue" (or silver)... was wrong: http://dailycaller.com/2016/05/04/7-times-nate-silver-was-hilariously-wrong-about-donald-trump/

His failure was simply his own personal bias. Ironman did a much better job than douche boy Nate.

joshuatrio says

Quit being so whiny and accept it.

I merely provided a post with some analysis. There was no whine there. If you want to hold up a straw man to argue against, be my guest. I agree that guru status is gone to the extent that it was ever there. I already explained why he got lucky in 2012. I guess it's not surprising that the same people who would attribute guru status in 2012 are let down. Maybe they'll apply the same failed logic and anoint guru status to someone else this year.

marcus is NOT deplorable says

EVENTS WITH A 30 to 36% CHANCE OF HAPPENING , CAN HAPPEN !!"

Those events happen whenever someone makes a 3 point shot. Most people just don't grok statistics and they can't learn to see past their incorrect intuition. That's why so many people say statistics lie or you can make statistics say whatever you want.

YesYNot says

Those events happen whenever someone makes a 3 point shot. Most people just don't grok statistics and they can't learn to see past their incorrect intuition. That's why so many people say statistics lie or you can make statistics say whatever you want.

Brah, whatever makes you feel better.

Nate was in the tank for Clinton. Five states wrong = fail.

Great job Ironman.

joshuatrio says

Brah, I understand that he gave Trump a chance.... like 29% (ish). Which is not even *GASP* 1 out of every 3 coin tosses!

YOU SEEM TO BE BLINDED by the fact that... he more or less predicted a Clinton landslide, not just with his model, but his own personal prediction, and has been laughing at Trump since the primaries.

OH, I just realized this is lying TPB (or someone just like him). You don't even follow 538, and never did. You hated him becasue he was saying your guy was unlikely to win. Again, for the entire 5 days before the election he had Trump at 34 to 36%.

Now after never reading his stuff even once you're an expert on what he does.

Go back to BReitbart. THey're geared more for your reading, intellectual honesty, and intelligence level.

Can polls be "wrong"?

marcus is NOT deplorable says

OH, I just realized this is lying TPB (or someone just like him). You don't even follow 538, and never did.

I've been on the site since 2006 (pre-forum/old forum).

Your wrong. Again.

joshuatrio says

One of the comments on poor Nate's site

Ugly crier

Ironman says

were giving Clinton a 4% to 8% lead based on that false reporting.

The rolling average and Nate's prediction based on the polls had her winning by 3.5%. She won by 0.2%. So, they were off by 3.3%. If they were oversampling by 5-14%, you'd expect them to be off by 2-11% if you think that there are really just 3% more dems than republicans. This is why I tried to get you to make a prediction on the popular vote before the election based on your argument. I gave a predicted popular vote and a range of likely results, and explained why. You did not provide a popular vote prediction and you did not provide a range. So now, you just have hand waiving.

Ironman says

YesYNot says

marcus is NOT deplorable says

EVENTS WITH A 30 to 36% CHANCE OF HAPPENING , CAN HAPPEN !!"

Somebody requote this chart for Marcus, because he won't see my post, as he has me on ignore. On Election Day (this chart), Nate had Trump at 28% and had Clinton winning with 302 in the EC... It was a MAJOR miss, and this was the trend with Nate the whole cycle:

-

-

-

"Expert Nate" was WRONG on the popular vote

He was WRONG on the E.C.

and he was WRONG on MANY states

TOTAL FAIL

Agree.

Ironman says

I posted my predicted chart of the states with my EC prediction in Thunder's thread, it's there too, go look at it. YOU never did that.

I never did it, because predicting state by state is a fools game, as I said beforehand. It's just too random. I asked you to predict the popular vote to see if you actually believed the argument you were making, and you wouldn't do it.

Ironman says

I was talking about ALL the polls, not just Nate's.

Nate doesn't have any polls. If the average of all polls oversampled by 9.5% (average of 5 and 14%), and there are only 3% more dems than republicans, then you would predict that the average of polls would be off by 6.5%. They were off by 3%, and it was for other reasons. The RCP rolling average Hillary advantage was 3.2% (Nate had 3.5%) and her actual advantage was 0.2% (http://www.realclearpolitics.com/epolls/2016/president/us/general_election_trump_vs_clinton-5491.html)

Some might find the chart in this article interesting. It shows that the accuracy of this poll (or lack there of) was completely normal relative to historical results.
https://www.washingtonpost.com/news/the-fix/wp/2016/11/10/how-much-did-polls-miss-the-mark-on-trump-and-why/

Ironman says

Nice try spinning it back to popular vote

Your point was about the popular vote being wrong, and that's all that's somewhat predictable so that's what I focused on before and after the election.

Look at the chart in the WaPo article. This result was within the expected range for anyone who was willing to look at the accuracy of polls over the years.

Ironman says

Predictable by what?? Your WaPo link says 1.7% (if you believe that). They missed Clinton by more than that and they certainly missed Trump by HUGE margin.

Go read the text. The 1.7 value is the 'candidate error,' which is defined as 1/2 of the difference between polled value and election value. If you prefer the raw difference, just multiply the y axis by 2. The data came from the National Council on Public Polls. Do you have reason to have an issue with them too?

Of course you can get a bigger error by averaging over a month. A month ago, Trump polled terribly, because his debate performance was fresh on everybody's mind. The last week or two before the election, he polled much better, b/c Comey reminded everybody about the emails and Trump managed to stay on message.

Wow, Bernie kicked the Shit out of Clinton in NH, MN, and IN. Very key states.

So sad the rigged DNC elections with Brazile giving debate questions (now more than twice, according to wikileaks which is still leaking Podesta emails) in advance to Hillary but not Bernie, and the Rigged Insider Superdelegates ended up nominating such a weak, despised candidate who drew millions upon millions less than Obama.

Whereas Bernie would have fired up the base and brought millions of new voters.

But it's all the electoral college, a known entity for 200+ years, that robbed Hillary.

Thunderlips Licks Shill Tears says

But it's all the electoral college, a known entity for 200+ years, that robbed Hillary.

Is this post in the right thread?

Ironman says

Are you going to whine on and on about Comey stealing the election from Hillary. You do know that MANY people vote early and when you're within the last two weeks, the majority of people have already made up their minds. There's very few undecided in the last weeks.

The point was not that Comey stole the election. It's that a 30 day average is not representative of what was happening during the week of the election. Hillary's lead was 5-6 points in the first two to three weeks, and 2-3 during the last week. Maybe it was Comey, maybe it was Trump, maybe it was the wind. But, a 30 day average is pointless when things change a lot during that period. You can clearly see this by going to the RealClearPolitics results. Stop pretending to be ignorant just because it favors your argument.

Are you going to recognize your mistake regarding 1.7 versus 3.4 or are you going to ignore that? History clearly shows that polls being off by 2-3 points is completely expected, and is definitely not proof of shenanigans by pollsters. Take your tinfoil hat off.

The problem was Trump was trending upwards before the Comey announcement.

Stop whinging. It was a shitty candidate that cost you the election.

Clinton was the strongest Democrat candidate though.

marcus is NOT deplorable says

At the 2011 White House Correspondents dinner, Seth Meyers and President Obama joked about the possibilityof a Trump White House run.

Elitism and Arrogance run common in personality traits.

Seth, Obama - Why dont you to form a business making something that is tangible and selling it. The rest of Americas had to do this to build our bridges, roads, plane, trains, technology. This building of things then gave people income to watch said Seth and TV, and also to elect Obama president.

In their arrogance, they became like fools, - entirely forgetfull of how they got to where they were.

America - Arrogance is wrong, humility is better.

Thunderlips Licks Shill Tears says

Wow, Bernie kicked the Shit out of Clinton in NH, MN, and IN. Very key states.

So sad the rigged DNC elections with Brazile giving debate questions (now more than twice, according to wikileaks which is still leaking Podesta emails) in advance to Hillary but not Bernie, and the Rigged Insider Superdelegates ended up nominating such a weak, despised candidate who drew millions upon millions less than Obama.

Whereas Bernie would have fired up the base and brought millions of new voters.

But it's all the electoral college, a known entity for 200+ years, that robbed Hillary.

What's really sad is that because of the near blackout by the msm regarding the wiki leaks emails, most Americans and nearly all overseas people have NO CLUE. And the the disingenuous democrat shills on this site are VERY happy about that and nearly entirely ignore the matter.

Personally if I was a Democrat, I'd be calling for Brazille and Podestas heads

Thunderlips Licks Shill Tears says

The problem was Trump was trending upwards

Thank you for proving my point that Trump trended upwards in the last week and a half, which means that Ironman shouldn't be using a month long average to get the polling result. Now maybe we can stop talking about Comey and go back to the point that the polling error was completely within the historical norm.

Peter P says

Clinton was the strongest Democrat candidate though.

False

If the polls were not the issue that leaves the cause as Muricans were fed up with Obama's BS?

There was low voter turnout but that is not the cause.

He needs to change name to Nate Aluminum, he was way off.