by Jonathan Hobratsch

In 2020, I accurately predicted the 2020 presidential election (aside from ME-2*) using an algorithm I created, which heavily favors momentum by state in presidential elections. Using the same algorithm, I predict that VP Kamala Harris defeats former president Donald Trump.

Donald Trump will become the first US president to be defeated as both an incumbent and former president.

Reasons why Harris likely wins:

• Harris and Trump are virtually tied in most of the battleground states, including the necessary Blue Wall States. Momentum has been on Harris’s side this final week, suggesting that late undecideds are possibly breaking for her in most of the battleground states, including the Blue Wall.
• Harris & Walz’s favorability is much better than Trump & Vance’s.
• The Direction of the Country polls have been slightly improving the last two weeks.
• The economic model that Nate Silver uses shows that the economy is “good,” but not “great.” Often it takes a below average economy to sink an incumbent party, although this is a weird election in which the incumbent dropped out. America has had the strongest COVID recovery of major nations in the world.
• Third Party voting is likely to be so low as to not really affect the election.
• Democrats lead in early voting in the Blue Wall States, especially in PA. More Democrats will vote on Election Day than in 2020 most likely because we COVID isn’t the same threat as it was in 2020. The demographic that doesn’t early vote often is the youth vote (18-29), and Harris has the lead there, according to polls.
• Conservative pollsters have been overproductive in this election, compared to 2016 and 2020. Partisan polling companies tend to favor their candidate. I seem to notice that the Democratic partisan polls (colored with a blue mark on FiveThirtyEight) drop about once a week, the Republican partisan polls (colored with a red mark) come out every day, sometimes three polls a day. This possibly means that Harris overperforms the polls, possibly by a lot, similar to how Trump did in both 2016 and 2020 (although he still lost this election).

Thanks for reading! Go vote!

* My algorthm missed ME-2 in 2020 primarily because there were so few polls there that it couldn’t really read it.