Nate Silver 538

The Juggler said:

MayDay10 said:

As stated, 538 recognizes how Quinnipiac skews and adjusts the probabilities.

Rasmussen... basically the Quinnipiac counterweight, has biden up by 3 in PA.  

there are a bunch in the middle, typically more accurate.  Their methods, trends. everything is taken into account.  Last turn, 538 had was one of the only places that gave trump a decent statistical shot.  

CORRECT.
AMUNDO.

Come on, Lex. What is so difficult here?

Suffolk today out with Biden up 6 in PA. "A" rated pollster. Given all the info we have, that sounds....SHOCKINGLY ACCURATE! lol

Filter all polling this month by those rated B and above (so you get rid of Rasmussen, Trafalgar and also Survey Monkey, etc)....and you have Biden up by an average of 7.8%

Looking good with under two weeks to go. Big debate tomorrow and we already know the president somehow made a fool of himself on 60 Minutes on Sunday and is doing nothing but drawing more attention to it. Get those confidence levels up, people.

mrussel1 said:

Lerxst1992 said:

The Juggler said:

MayDay10 said:

As stated, 538 recognizes how Quinnipiac skews and adjusts the probabilities.

Rasmussen... basically the Quinnipiac counterweight, has biden up by 3 in PA.  

there are a bunch in the middle, typically more accurate.  Their methods, trends. everything is taken into account.  Last turn, 538 had was one of the only places that gave trump a decent statistical shot.  

CORRECT.
AMUNDO.

Come on, Lex. What is so difficult here?

Suffolk today out with Biden up 6 in PA. "A" rated pollster. Given all the info we have, that sounds....SHOCKINGLY ACCURATE! lol

Filter all polling this month by those rated B and above (so you get rid of Rasmussen, Trafalgar and also Survey Monkey, etc)....and you have Biden up by an average of 7.8%

Looking good with under two weeks to go. Big debate tomorrow and we already know the president somehow made a fool of himself on 60 Minutes on Sunday and is doing nothing but drawing more attention to it. Get those confidence levels up, people.

I think my point is simple and I hope it’s clear I believe Biden has a better chance to win than trump. But I don’t think RCPs data can be disregarded because we think Nates process is the best. He was actually outperformed by RCPs much simpler process last time. I don’t believe that guarantees anything this year, but gives me just enough pause to say an 88% race should not have tipping point states within or close to the MoE.

What Nate said last time, “trump is one poling error from winning” seems to be applicable again. Last time nationally. This time in PA FL NC AZ which are the key tipping point states all within MoE. 

RCP doesn't do anything with the polls but average them, right?  It's not an analytical exercise. 

MayDay10 said:

What does ICP have?

Lerxst1992 said:

The Juggler said:

And just like that after some higher quality polls yesterday, RCP is back to about a 5% lead in PA while 538, who doesn’t sway much with garbage polling, remains around 6%. 

I’ll make the point again, in 2016 RCP was more accurate than 538. In 2012, it probably was not. But that’s not enough to pre judge the current data.

I am liking most of the PA polls yesterday, except that Fox one at +5. As you know, Fox is a strong lean D, per  Nate (+1.5 IIRC).+5 Biden in a poll with a strong D lean seems to be within MoE. Hopefully pollsters are overcompensating for 2016 turnout, and it’s more like 2012. But that’s hope, not data.

benjs said:

Lerxst1992 said:

The Juggler said:

And just like that after some higher quality polls yesterday, RCP is back to about a 5% lead in PA while 538, who doesn’t sway much with garbage polling, remains around 6%. 

I’ll make the point again, in 2016 RCP was more accurate than 538. In 2012, it probably was not. But that’s not enough to pre judge the current data.

I am liking most of the PA polls yesterday, except that Fox one at +5. As you know, Fox is a strong lean D, per  Nate (+1.5 IIRC).+5 Biden in a poll with a strong D lean seems to be within MoE. Hopefully pollsters are overcompensating for 2016 turnout, and it’s more like 2012. But that’s hope, not data.

A broken clock is right twice a day.

Lerxst1992 said:

Jearlpam0925 said:

benjs said:

Lerxst1992 said:

The Juggler said:

And just like that after some higher quality polls yesterday, RCP is back to about a 5% lead in PA while 538, who doesn’t sway much with garbage polling, remains around 6%. 

I’ll make the point again, in 2016 RCP was more accurate than 538. In 2012, it probably was not. But that’s not enough to pre judge the current data.

I am liking most of the PA polls yesterday, except that Fox one at +5. As you know, Fox is a strong lean D, per  Nate (+1.5 IIRC).+5 Biden in a poll with a strong D lean seems to be within MoE. Hopefully pollsters are overcompensating for 2016 turnout, and it’s more like 2012. But that’s hope, not data.

A broken clock is right twice a day.

Exactly - same goes for an some guy I work with who now follows only Rasmussen because they were right in 2016. I proceeded to tell him every squirrel finds a nut. He also went on to think CA had a chance to go red because of some dumb tweet, sooo.... 

Confirmation bias is a helluva drug.

It’s more than just confirm bias or a broken clock. NCWs are the largest demo that most pollsters missed, and nobody is sure if they will bring a bigger tidal wave of support this time.

I also think its possible the AA dem base will also show up in big numbers, with Obama helping. I certainly miss having a leader like him. But that demo is much much smaller, so polling error there has a lower impact