2016 Postmortem
In reply to the discussion: So. FL, MO, NC, OH, and IL vote on Tuesday. My predictions. [View all]Jarqui
(10,131 posts)I do not think anyone should mock you given how far off the polling has been in these primaries. It's not easy and we're all having to guess some because the information is so scant or poor quality.
i think Nate is really sticking his neck out asking for trouble with these >98% and >99% predictions - the polls have not been that good. Just on that alone, I don't think he should go that high. I suspect we'll see an article before the primaries are out on how he's "adjusting his modeling in primaries" to deal with this. The odds are in his favor that it will work out as he says but I think the polling data is so suspect that the chances of him being wrong is greater than a fraction of 1%.
The pollsters claim their margin of error is +/- 5% or so but the polls of polls is often out closer to 10 points in the south and was off by 22 pts or so in Michigan. That is messed up. That is certainly not something we see in the general election. +10% or +12% is typically "safe" or ">99%" but in these primaries +/- 10-20% seems to the margin of error (which is almost nuts)
So thank you for sticking your neck out a little.
If Bernie can keep the losses down, I think the demographics turn for him after Mar 15th and it could get very interesting.