US 2020 Presidential Election Predictions

This is the only site with all 2020 US Election Predictions

Toss-Ups Analysis

All the forecasters are looking at the same state polling data. However, they interpret the data differently. How far back does one go in averaging the data? Surveys may contain 500 to 1500 respondents, a very meager sample in states like Florida, where 9 million people voted in the last election. And how much should they factor in historical trends in ranking the states? And they can also factor in recent elections for Senators, Representatives and Governors.

Polling data provide an estimate of the percentage of voters which might vote for a candidate. The key statistic is the vote margin in the sample poll. The website "Election Projections" uses the following race ratings: “Weak” – under 5% “Mod” – 5% to under 10% “Strong” – 10% to 15% “Solid” – over 15%. Of course, knowing that Biden will win California, for example, with a margin of 15% or higher still does not translate directly into the chance of a Biden win in California, except the value should be close to 1. A few recent surveys with margins above 5% can put a candidate in the lean or likely group instead of as a toss-up.

In my last comment, I suggested four core toss-up states, PA, AZ, WI and NC which spell PAWN. I realize that many forecasters include Nebraska Congressional District 2 (aka NE2) with one electoral vote as a toss-up. So, our new PAWN includes NE2, which puts 57 electoral votes in the toss up group.

Sabato's Crystal Ball toss-up's consists only of the new PAWN group with 57 EV's. Florida and possibly Michigan qualify as consensus toss-ups. Trump won Florida by a very narrow margin in 2016. Consistent with Cook's Report, our consensus toss-ups are now PAWN + FL +MI. Almost forgot NE CD2 is a core toss-up too. This brings the toss-up pool up to 106 EV's.

PAWN = PA, AZ, WI, NC and NE CD2 for 57 EV's.

Real Clear Politics (RCP) considers 19 states/entities to be toss-ups for a total of 230 EV's. Perhaps not perfect toss-ups, but they are definitely flippables or states where we could definitely see surprises in November. This is my consensus look at the states, with light red and blue being tilt or lean and it includes all 19 states from RCP.

Way back in January 2020, when there were about 5 likely Democrat candidates, Skylar Dale used an analytical approach to predict the percent of votes each candidate would receive. His percentages are shown for the Democratic party. You can see the easily see the similaries among these forecasts.

Texas (46%)’, ‘Iowa (46%)’, ‘Ohio (46%)’,‘Georgia (47%)’, ‘Arizona (48%)’, ‘North Carolina (49%)’, ‘Florida (49.37%)’, ‘Wisconsin (50.02%)’, ‘Michigan (50.39%)’, ‘Pennsylvania (50.41%)’, ‘Minnesota (50.89%)’, ‘New Hampshire (51%)’, ‘Maine (51%)’, ‘Nevada (52%)’, ‘Colorado (53%)’, ‘Virginia (53%)’

He predicts a very close race with Biden winning with 279 EV's. The poll based predictions are pretty much higher than this. But look at how much has changed in the race since January. Forecasters are moving some states further to the "lean Democrat" group. And we have around 5 months to go.

See Link: Skyler Dale, Medium.com

David Lord