cities. physics. food. environment. fatherhood.
Random header image... Refresh for more!

Elections return, with a spreadsheet

Sunday’s Super Bowl was an interesting game: the lead changed several times and until the very last seconds of the game, it seemed like either team could win.

Emotionally, at least, there’s a similarity between watching election night returns and watching sports: as the votes tally up, one can form a mental picture of a literal race, and if your favored team or candidate is behind, you cheer when the gap closes. Of course in sports, the actions of the players determine the course of the game, and crazy things can happen. In elections, it’s all over once the polls close. Barring irregularities like the 2000 presidential election in Florida, it really doesn’t matter what order the votes are counted in, and once the votes start to be counted, there is nothing anyone can do to get more votes. Cheering doesn’t give anyone a boost.

Of course, for all those involved in a political campaign, it also ends once the polls are closed, especially for the losing candidate. But even for the winning candidate, the dynamic of everyone involved changes dramatically. Those hours of uncertainty, after the polls close but before the winner is known, are the only possible time to have one last gathering of the campaign, and it might as well be a party, and you might as well find out how you’ve done.

And for everyone at home, watching the election returns can be entertaining, to know as soon as possible what happened. So as thoughts of Super Bowl turn to thoughts of Super Tuesday, I present my rudimentary election watching spreadsheet.

The television networks often project a winner even when it’s mathematically possible for either candidate to win, and the spreadsheet I offer here lets you play along, too. It only uses three pieces of information: the number of votes each candidate has, and the percent of precincts that have reported.

If we make the approximation that all precincts will have the same number of voters–not generally true, but hard to get a better number without detailed precinct-by-precinct data–then we can project the total number of votes that have been cast, and from that calculate the number of votes remaining to be counted, and of those the number each candidate would need to win, and finally, what percentage of the remaining votes each candidate would need.

These percentages are really illuminating: if you calculate that a candidate who has 45% of the vote so far would need to have 67% of the remaining vote to win, then you could call the election for the other candidate with a fair degree of confidence.

To use, just put the most recent vote counts in cells A3 and B3, and the percentage of precincts reporting in C3. The rest is calculated automatically. Use the Fill Down command to create multiple lines for running progress.

Watch Returns

3 comments

1 Ken Monahan { 02.05.08 at 4:58 pm }

The iMac in the Bahrain Airport won’t let me download your spreadsheet. : (

2 thm { 02.06.08 at 10:21 pm }

Hmm. It’s a really simple set of calculations, so it shouldn’t be too difficult to write as a Javascript or something; perhaps I’ll get something together before the General election.

Then again, using my spreadsheet I had convinced myself last night that Clinton would get more of the popular vote in Missouri than Obama would–with 91% of precincts reporting, Obama had about 49% of the vote and needed to get 57% of the remaining vote… which he did.

3 Ken Monahan { 02.15.08 at 3:57 am }

I don’t think it was the spreadsheet, I think it was just the security on the computer I was using.

Leave a Comment