Here we go – if I stick to my once a month poll review, there will only be one more just before the election.
| STATE | 2020 RESULT | Silver Bulletin (September 1) | Silver Bulletin (October 1) | 538 (October 1) | RCP (October 1) |
| Arizona | Biden +0.4% | Trump +0.6% | Trump +1.5% | Trump +1.5% | Trump +2.1% |
| Georgia | Biden +0.3% | Harris +0.9% | Trump +1.0% | Trump +1.3% | Trump +1.5% |
| Wisconsin | Biden +0.6% | Harris +3.2% | Harris +1.9% | Harris +1.6% | Harris +0.6% |
| North Carolina | Trump +1.3% | Trump +0.4% | Trump +0.5% | Trump +0.7% | Trump +0.7% |
| Pennsylvania | Biden +1.2% | Harris +1.3% | Harris +1.2% | Harris +0.6% | Trump +0.1% |
| Michigan | Biden +2.8% | Harris +1.9% | Harris +2.1% | Harris +1.9% | Harris +1.4% |
| Nevada | Biden +2.4% | Harris +0.9% | Harris +1.8% | Harris +1.0% | Harris +1.1% |
The first thing that stands out is there is no disagreement between the weighted poll averages (Silver and 538) on the more likely winner of each state. They also agree with the RCP unweighted average with the exception of Pennsylvania. Pennsylvania just looks dangerously close, with a 1% polling bias towards Harris (vs. how people in the state end up voting) making it a tossup. Still, you would rather have that polling average in your favor than against you. The difference between the weighted averages and RCP suggest that there either a lot of polls of Pennsylvania voters that the weighters consider Republican-biased garbage, a big recent trend toward Harris in Pennsylvania (because they rate more recent polls higher), or a combination. I can tell you from personal experience that the Democratic get-out-the-vote operation in my home city of Philadelphia is in hyperdrive, but I also assume the Republicans equivalent is in hyperdrive in Republican-leaning counties (like my old home county of Luzerne in Northeast Pennsylvania.)
If the polls are reasonably accurate, Georgia and Arizona might be moving out of Harris’s reach, and it is hard to believe North Carolina and Georgia are that culturally different (think about the Charlanta mega suburban sprawl cluster-f which is basically one thing).
If Harris wins Pennsylvania, she seems likely to win Wisconsin and Michigan and the electoral vote as a whole. Nevada would pad the score a bit for Harris, but it would not offset the loss of Pennsylvania.
Polymarket gives Harris 50% to 48% odds. Predict prices her at 56 cents to Trump at 48 cents, with other candidates given about 8 cents.
So all the signs kind of point to Harris, but if there is a systematic error of 1-2%, Trump could still pull it out.