The numbers I'm seeing in media and academic polls seem to be remarkably non-reflective of what I assume to be popular displeasure in states that have suffered from riots and excessive lockdowns. Now I realize this time of year the polls are largely designed to demoralize the opposition and boost your troops. but I'm starting to wonder if the demise of the land line, the rise of the shy Trump voter and the fragmentation of the media landscape has deprived them of any statistical validity at all. That certainly seemed the case in 2016 and I have to assume that's just as true now. And I suspect that the 10 states which will decide this election have been thinly polled. A sample size of of 600-800 people in a good sized Midwestern state seems small to me, given the above. particularity when party composition estimates are applied to the sample.

So for those of you who are familiar with the polling the candidates themselves use, is there any confidence in there predictive ability or are they just as inexact as the media and academic polls?

P.S. Jeff Gundlich said yesterday he thought Trump would win and I have to say it looks to me like the markets think the same thing. if that's a correct reading (i.e. I think markets are smarter than pundits), do the professional polls reflect this?

The numbers I'm seeing in media and academic polls seem to be remarkably non-reflective of what I assume to be popular displeasure in states that have suffered from riots and excessive lockdowns. Now I realize this time of year the polls are largely designed to demoralize the opposition and boost your troops. but I'm starting to wonder if the demise of the land line, the rise of the shy Trump voter and the fragmentation of the media landscape has deprived them of any statistical validity at all. That certainly seemed the case in 2016 and I have to assume that's just as true now. And I suspect that the 10 states which will decide this election have been thinly polled. A sample size of of 600-800 people in a good sized Midwestern state seems small to me, given the above. particularity when party composition estimates are applied to the sample.

So for those of you who are familiar with the polling the candidates themselves use, is there any confidence in there predictive ability or are they just as inexact as the media and academic polls?