Reuters/Ipsos Natl: Trump +2

[SherlockHound]

The difference between 2012 and 2016 is that the people decided on Romney or Obama because they liked the candidates and their political stances, while this election is all about chosing the lesser of two evils.

Reminds me of 2004, Bush v Kerry.

[Ebsy]

Ah so I was right when I said that they changed their methodology/screen. The swings were too massive to be real.

[Erich Maria Remarque]

Very intresting, indeed!

Shy Trump. As I said, allmost the right wing parties in Europa were underestimated by polls. When the Election Day comes, they usually increase with 2-4  pps. And BREXIT as well

https://en.wikipedia.org/wiki/Opinion_polling_for_the_United_Kingdom_European_Union_membership_referendum#2016

Brexit polls during the last week before referendum:
Tie
Brexit -1
Brexit -3
Brexit -7
Brexit +2
Brexit -1
Brexit +2
Brexit -6
Brexit -8
Brexit +1
Brexit -3
Brexit -2

And last poll, one day before referendum:
Brexit -10

Reality? Brexit +3.78%.
All the polls underestimated Brexit.

[pbrower2a]

There is no typical Presidential election.

It is possible to bunch them by scale of victory, so elections of 1972 and 1984 are extremely similar , and those two have an obvious parallel (if an inversion of Party) in 1936.

When the victor is the same in two adjacent elections the maps could be very similar -- as 2012/2008; 2004/2000; 1996/1992; and of course 1956/1952.

One election has a near-mean in percentages of popular and electoral votes -- 2012.  But elections are otherwise close or they diverge as several levels of landslide. They do not bunch around 62% of of the electoral vote as 2012 was; they clearly diverge, getting much closer or getting farther away from closeness.

The 2012 election was a freak even if it got a mean-like result.  As it is I can predict three possibilities for 2016: a close election with either Trump or Clinton winning or a blowout victory for Clinton.  

[Wells]

Rasmussen shows Clinton leading.
Reuters shows Trump leading.
Junk.