Count every vote. #CountEveryVote
— Donie O'Sullivan (@donie) November 3, 2020
This simulation can make you either reassured or completely crazy:
— Nate Silver (@NateSilver538) November 2, 2020
How this works: We start with the 40,000 simulations that our election forecast runs every time it updates. When you choose the winner of a state or district, we throw out any simulations where the outcome you picked didn’t happen and recalculate the candidates’ chances using just the simulations that are left. If you choose enough unlikely outcomes, we’ll eventually wind up with so few simulations remaining that we can’t produce accurate results. When that happens, we go back to our full set of simulations and run a series of regressions to see how your scenario might look if it turned up more often.
In simplified terms, the regressions start off by looking at the vote share for each candidate in every simulation and seeing how the rest of the map changed in response to big or small wins. So let’s say you picked Trump to win Texas. In some of our simulations, Trump may have won Texas very narrowly and also have narrowly lost some toss-up states. But in simulations where he won Texas by a big margin, he may also have won big in toss-up states and pulled some Democratic-leaning states into his column, while the reverse may be true in simulations where he lost the state. We figure out how every other state tended to look in that full range of scenarios, tracking not just whether the candidate usually won other states but also how much he generally won or lost each one by.
After all that, we take some representative examples of scenarios that include the picks you made and use what we learned from our regression analysis to adjust all 40,000 simulations, and then recalculate state and national win probabilities. Finally, we blend those adjusted simulations with any of the original simulations that still apply and produce a final forecast.
*Maine and Nebraska award two electoral votes to the winner of the statewide race and one electoral vote to the winner of each congressional district.
This year, polls show President Trump may lack one of the key ingredients that helped him engineer a shocking upset in the Electoral College in 2016: undecided voters who broke late to him.
- Many of these voters swung to Trump in the final days of the election: a study conducted by the American Association for Public Opinion Research found that approximately 13% of voters in Wisconsin, Florida and Pennsylvania made up their minds about who to vote for in the final week before Election Day, the Wall Street Journal reports.
- These voters backed Trump by almost 30 points in Wisconsin and 17 points in Florida and Pennsylvania, according to the study, meaning the president did more than just outperform polls.
- This year, polls show there are fewer undecided voters, and most of them appear to be leaning towards Joe Biden, not the president, which reduces the likelihood for a last minute shift in the race in Trump’s favor.
- A new Wall Street Journal/NBC News poll of voters in 12 battleground states—Arizona, Florida, Georgia, Iowa, Maine, Michigan, Minnesota, Nevada, New Hampshire, North Carolina, Pennsylvania and Wisconsin—shows the race basically unchanged in the last week, with Biden ahead by a combined 5 points (51%-46%), after leading by a combined 6 points in the same poll one week earlier.
— Olivia Nuzzi (@Olivianuzzi) November 1, 2020
— Ben Collins (@oneunderscore__) October 29, 2020
— MIT Technology Review (@techreview) November 2, 2020
— MSNBC (@MSNBC) November 2, 2020
— Jason Campbell (@JasonSCampbell) November 2, 2020
— It's Going Down (@IGD_News) November 3, 2020
— Ben Meiselas (@meiselasb) November 3, 2020
— New York City Antifa (@NYCAntifa) November 3, 2020