I first published this little app two years ago, but I first used it four years ago to figure out that Donald Trump was winning a lot earlier than did many of the TV pundits. Click here to run the app and bookmark it to use as the election returns roll in. I have improved it this year to work better on phones and tablets.

A common scenario is that your favored candidate is down by a few points, with some of the vote still not counted, or perhaps the candidate is leading by an uncomfortably small margin. The key question is how much of the remaining vote your candidate needs to get in order to finish “first past the post.” If the “remaining vote percentage” number gets unreasonably high, as it did far too early in key states in the Clinton/Trump race, then you will know whether to make it an early night. If the number still looks achievable, then you will want to hang in for some better data.

It is not a hard calculation, but the answer can be counter-intuitive, and it seems that very few of the “talking heads” highlight this number. By the way, for a longer discussion of my thoughts on the art of election forecasting, here is link to an earlier post on “Some caveats about election statistics.”

In the case above, my candidate is trailing, winning just 45% of the votes cast, but only 30% of the vote has been counted. In this case, there are no third-party candidates, so the percentage needed to win is 50% plus one. We then need, as the app shows, to get 52.14% of the remaining votes in order to end up with 50% of the total vote. As more of the vote comes in this number begins to rise precipitously:

Our odds are much better if we are tracking at 49%, but once we pass 90% of the votes counted the win still quickly tracks out of reach:

You can also use this tool to determine the safety of your margin if you are currently ahead. If your candidate currently has earned 52% of the vote, and 80% of the vote has been counted, he or she needs only 42% of the outstanding votes to still win.

If there are third party candidates, you can take the percentage needed to win down by half of the effect of the total third-party vote percentage (e.g., if third-parties are getting 2% of the vote, then you can probably win with just over 49% of the vote).

Note also that some state constitutional referenda, such as in Florida, require a 60% plus one vote in order to pass, so just change the “percentage-to-win” value to 60 for these ballot issues. And some state elections, such as Georgia’s Senate races, require 50% of the vote plus one to avoid a run-off election.

The record-setting number of mail, absentee and early votes cast already in the 2020 presidential election requires that we pay special attention to that “percentage of the vote in” number. In the past, this was often presented as a percentage of voting precincts reporting, which makes no sense this year. NBC, for one, has stated that they will be reporting their best estimate of counted votes as a percentage of votes actually cast at various points through the tally process.

Happy counting and good luck!