When will enough votes be counted in each state to determine who the winner is? And how will that affect when we know the results of the election? Will the early vote even matter in some states?

I did some nerdy spreadsheet math and below you will find the order in which I predict the states will be called on election night for the presidential race. Specifically, the predictions are for when the associated press will call a winner for each state. There might still be ballots being counted, but the margin will be wide enough to call the winner in these states at these times.

The methodology involves several variables. I used probabilities of winning each state from fivethirtyeight.com and the pre-processing, tabulation, and deadline data for each states’ early voting from the New York Times. I also used the timezone data for each state from ballotpedia.org. Further methodology notes are below.

Based on my model, I predict that we will know who the president-elect of the United States is at 1:06pm on November 4th, when Pennsylvania decides that their margins are wide enough to call their statewide winner. Assuming that the order of states is approximately correct, most likely scenarios (~96%) show that November 4th should be the day that we know who won, with a small chance (~4%) that the results will be learned on November 6th, and a vanishingly small chance (<1%) that we find out on November 3rd.

The model predicts that there will be several states that are not called until after November 3rd. Many states are predicted to be called on November 4th—most notably the swing states of Florida, Virginia, Wisconsin, Pennsylvania, Minnesota, and Michigan. Alaska is predicted to be called on November 5th, and Iowa and North Carolina are predicted to be called on November 6th. Maine’s 2nd district is expected to be called on November 30th.

My model also predicts that there will be two states in which the election is too close to call across the entire month of November, thanks to the margin of victory being very close, and disputes about voting.

Kentucky 7:00pm

Vermont 7:00pm

Connecticut 8:00pm

Delaware 8:00pm

Illinois 8:00pm

Maryland 8:00pm

Massachusetts 8:00pm

Oklahoma 8:00pm

Rhode Island 8:00pm

Washington DC 8:00pm

Indiana 8:22pm

South Carolina 8:38pm

Arizona 9:00pm

Colorado 9:00pm

Louisiana 9:00pm

Nebraska Statewide 9:00pm

Nebraska 1st 9:00pm

Nebraska 2nd 9:00pm

Nebraska 3rd 9:00pm

New York 9:00pm

North Dakota 9:00pm

Texas 9:00pm

Wyoming 9:00pm

Maine 1st 9:13pm

Tennessee 9:17pm

Missouri 9:35pm

Maine Statewide 9:35pm

Arkansas 9:43pm

Montana 10:00pm

Nevada 10:00pm

South Dakota 10:14pm

New Mexico 10:15pm

Alabama 10:41pm

New Hampshire 10:47pm

California 11:00pm

Hawaii 11:00pm

Idaho 11:00pm

Washington 11:00pm

Oregon 11/4/2020 12:14am

Florida 11/4/2020 1:50am

Virginia 11/4/2020 2:06am

Wisconsin 11/4/2020 2:44am

West Virginia 11/4/2020 5:26am

Kansas 11/4/2020 5:37am

New Jersey 11/4/2020 6:46am

Mississippi 11/4/2020 12:04pm

Pennsylvania 11/4/2020 1:06pm

Utah 11/4/2020 2:02pm

Minnesota 11/4/2020 5:21pm

Michigan 11/4/2020 5:43pm

Alaska 11/5/2020 3:33pm

Iowa 11/6/2020 1:17am

North Carolina 11/6/2020 4:47am

Maine 2nd 11/30/2020 7:59pm

Ohio: Too Close to Call; decided in court

Georgia: Too Close to Call; decided in court

Based on the NYT data, I calculated whether each state would be likely to be finished counting enough absentee ballots on November 3rd to help show a clear winner. For states with tight deadlines, I assigned penalties to the hour that the state is predicted to be called based on when ballots were preprocessed and tabulated, and when they needed to be received by. Those penalties when then multiplied based on the odds that the early vote would matter in calling each state; closer elections were assigned higher odds of mattering. For states with >99% chance of a particular candidate winning, no delay penalty was assigned. For states considered “swing states” in this election cycle, the penalty was given a “Tossup bonus careful counting” multiplier to reflect a slowdown or various swing state shenanigans that make it harder to count. In states where the early votes could only begin being tabulated on election day itself, a further penalty multiplier (Same Day Absentee Counting Penalty) was applied, due to the amount of time needed to actually count the ballots. Then, using the timezone data, I adjusted the expected time for each state to be called based on the above penalties, keeping in mind that the states cannot be called before their polls close on election day.

References:

https://ballotpedia.org/State_Poll_Opening_and_Closing_Times_(2020)

https://www.nytimes.com/interactive/2020/10/13/us/politics/when-votes-counted.html

https://projects.fivethirtyeight.com/2020-election-forecast/