US 2020 Presidential Election Predictions

This is the only site with all 2020 US Election Predictions

Update on Election Forecasts: October 25, 2020

Eight days to go. We continue to add to forecasts both on the home page and the non-polling based methods page. In this home stretch, both Trump and Biden are pushing home their messages in the well defined 10 swing states as listed at the bottom of this blog. If there is any doubt these 10 states will define the election, just follow where the candidates will be speaking this week.

Poll based forecasts:

My way of splitting up toss-up votes caused unexpected noise in the forecast. So, I adopted the uneven split of toss-ups as I explained in an earlier posting. No one has to accept my toss-up split. I've listed out the predicted EV''s excluding the toss-ups, and they all point to a victory for Biden.

The result is that have 15 forecasts which all predict Biden will win with 334 to 335 EV. There are seven forecasts which predict Biden will receive between 341 to 389 electoral votes. They may be right, because they are including the 4 states which Trump holds a razor thin margin: Georgia, Iowa, Ohio and Texas. One or all could come down on Biden's side. See discussion at the end of this blog on the "GOTI's.

Non-Poll based Methods:

For the "non-poll based methods" I have 11 new forecasts, from papers submitted to the American Political Science and Politics Association and posted on the Cambridge University website. Three of the papers which predict Trump will win. The average electoral votes for those predicting Biden as the winner is 333 EV. Is it right to average only Biden winner papers? Maybe not.

Betting Sites: (corrected Oct 26)

I like to check with the gamblers who back their forecasts with their own money. The Predictit website has Biden leading by 290 EV's. with both Florida and North Carolina now in the lean Republican camp. Arizona (11) could go next and drop Biden score to 279 EV's. Real Clear Politics summary of betting sites, gives Biden chances of winning at 66%, which is below the probability of winning from a number of websites, including JHK forecasts and the 538 website.

Sim 538 forecast:

The Sim 538 forecast predicts Joe Biden will win exactly 334 electoral votes. I remain convinced that Trump has a very reasonable chance to win the election. With my uneven split of toss-ups, I am presuming that North Carolina and Florida go to Biden, with 44 electoral votes. But if he loses these two states, it becomes a very close race, with Biden in the lead at 290 EV, just 21 above the magical 269 line.

Evaluating the forecasts:

Once the dust settles, and hoping it does soon after November 3, I will provide a review of the forecasts. I will base scores on the election outcomes first in terms of electoral votes, and then each state's outcome. I will have to somehow penalize those forecasters who label many states as toss-ups.

So, this is really where we stand to now. The Sim 538 prediction of a Biden win with 334 electoral votes is final. In the end, I am joining the crowd. My forecast considers the entire GOTI gang as Republican. We shall see.

A Podcast from Fivethirtyeight:

Iowa, Ohio, Georgia and Texas polling data show only very thin margin for Trump and many forecaster consider them as toss-ups. A very good question, is what kind of voter has Joe Biden been able to attract, that Hillary Clinton could not. Obama won Iowa and Ohio in 2012. The answer may be a bit surprising, because the idea of the Trump supporter as white and lacking a college education is highly over generalized. There is an excellent podcast on this topic on the fivethirtyeight website. See the link below:

The most competitive states are not Michigan, Wisconsin or Pennsylvania.

David Lord

October 25, 2020

10 Swing States Abbreviations, Name and EV's

MI = Michigan (16)

PA = Pennsylvania (20)

AZ = Arizona (11)

WI = Wisconsin (10)

NC = North Carolina (15)

FL = Florida (29)

TX = Texas (38)

GA = Georgia (16)

IA = Iowa (6)

OH = Ohio (18)

Congressional Districts:

NE2 = Nebraska Second Congressional District (1). Nebraska has 3 CD's each with 1 vote. The state has 2 votes, for a total of 5 votes.

ME2 = Maine Second Congressional District (1). Maine has 2 CD's, each with 1 vote. The state has 2 votes for a total of 4 votes.