Ready The Clown Car: The First Batch of Democrats Are Ready To Announce Their 2020 Bids

trump?

 

Yea, I guess I remember the predictions because they predicted an absolute slam dunk for Clinton.

I guess I naively assumed those predictions bore some reality based on polling... I guess you’re telling me those predictions were made up out of whole cloth by the left wing media?

 

Actually, the polling was accurate, or rather, it was no more or less accurate than other elections in recent history. States, individual districts, the national popular vote- most things were within the margin of error.

The projections that certain sites made were based on statistical modeling made with the polling data- many of them were far off, but the error is easier to understand if you know that winning the EC while losing the popular vote is very hard to do. And the higher the popular vote margin loss, the harder it is to pull the EC victory.

Trump losing the popular vote by almost 3% while winning the EC was an incredibly narrow victory. In an election with, what, 120-130 million votes, a combined total of 75k votes spread across WI, MI, and PA decided the EC victory. That's grabbing victory with one millimeter of one fingernail.

Most organizations that made their own projections leaned heavily on popular vote, and with HC polling 4-5% victory there (which was within the margin of error), they were projecting %90-99 percent chance of victory.

Except fivethirtyeight.com.

As much as conservatives like to **** all over Nate Silver and 538, it was his model that had Trump gaining ground in the final week (especially after Comey did his thing. Thanks, Comey) and by election day 538 had a projection with Trump at approx %33 chance of victory. Also, a couple of weeks earlier, Nate wrote an article analyzing a weakness in the so-called "blue wall." Turns out his warning was dead-on.

538 put a ton of work into their models, projecting demographics and turnout for every single district, weighting individual polls based on past performance, and running a thousand election simulations every day and posting their results.

Y'all are really missing out on some great fact-based reporting when you keep blasting all media outside of the right wing.

 

Yeah it really is a lot like the Republicans in 2016.

 

The polls really were not “so wrong”, the outcome was within the margin for error.

 

The electorial college system means that if a candidate wins a state by .1% point, that candidate takes 100% of the electoral votes for the state.

So even if Hillary had a much higher probability of winning, all it took was a couple of swing States where Donald won by less than a percentage point for him to win the electoral college vote.

Theoretically a candidate could win all 538 electoral votes which would seem like an epic landslide but only win by .1% pt in each state except in Nebraska and Maine where they split electoral votes but they account for a small amount anyways.

 

This is one of those myth that continue and will continue unfortunately, I think largely because people are confuse about what is a poll and what is a prediction.

Polls are simply a snap shot in time that reflect that exact moment(s). It's always backward looking and never forward looking. Prediction is forward looking based on some statistic models, some of which might be good and some are complete garbage (those that include personal biases of the people creating the models). People, companies, websites that uses prediction rarely explain or offer up their model - some see it as their "secret" sauce and they wouldn't share.

The polls weren't so wrong in 2016. It's about the same or even performed better than 2012 for the most part. What was wrong were the talking heads, the analysis, those people with loud mouth and loud opinion making predictions - this include newspaper and sites that make their own predictions. They often use the word "poll" adding to the confusion of what is a poll and what is a prediction. And what was also wrong was right after the election, the same set of people that were wrong, were again wrong about the polls - that it was the polls not them and their predictions that were wrong.

So, no the polls weren't so wrong. They actually have probably gotten better over time with better tools. The politician will continue to use them, knowing well that it is a good tool.

Prediction will get better over time, but it's still early and since again, the methods are not shared, most are on their own to learn from past mistake, to make improvement and so I think it will take a while for it improve.

 

If the EC states winner takes all, then why is Nebraska and Maine the exception?

 

Because it's up to the state to decide?

 

Ah ok. Your first sentence implied the EC determined winner takes all. If you prefer the split system, perhaps you should move to Nebraska or Maine. Or petition your state to change their rules.

 

My post wasn't a desire. It was an explanation of why a prediction can be wrong with accurate polling data within the margin of error.

Though I prefer electoral votes to be spilt in each state, it isn't going to influence my decision of where I want to work and live.

 

This. If polling is bullshit, then gerrymandering would be a useless waste of time.

I think the disconnect is that the media allowed polls with suspect sampling methods into the narrative which created an incredibly false degree of confidence among it's readers and public opinion.

The other unmentionable is that the Trump campaign had access to Hillary's and the DNCs polling and voter operations. This is a verified fact, even if collusion is not. Its the literal national playbook of the entire campaign, which means that even if Trump had a shoestring budget with a Mickey mouse operation like many of his adversaries were claiming at the time and many of their own admitted in hindsight, they still had a field of vision far better than Hillary.

 

These are novel concepts. Let him ease into them.

 

So, if I give the Rockets a 98% chance of winning the first round of the playoffs, and then they lose, was my prediction wrong or did we just actualize the 2%?

 

i have no idea why some think Trump is a lock as a candidate.

he won a fluky election in 2016 and by every metric he is less popular now and shows no signs of trying to get Independent voters in board again.

not saying he can't win but nobody should be that confident about it.

 

Or maybe the predictions/stories everyone was being fed about the Clintons assured victory were contrived and weren't worth the paper they were written on? Much better chance of that than the 1.9% being realized.

https://www.vox.com/2016/10/18/13311128/polls-trump-clinton-2016-winning

Seems legit.