Electoral-Vote.com (formally Electoral Vote Predictor) is a website created by computer scientist Andrew S. Tanenbaum. In the periods leading up to U.S. federal elections, the site's primary content is poll analysis to project election outcomes. Since the 2016 elections, the site also has featured daily commentary on political news stories.
The site was created during the lead-up to the 2004 U.S. Presidential election to predict the winner. The site tries to improve on national polls usually reported in the media, by instead analyzing the most recent polls on the state level, thus simulating the process by which Electoral College votes are determined in the actual election. Updated throughout the campaign, visitors can see who is "ahead" at any time.
Instead of the site lying fallow between election cycles the way it did before, starting in 2015 the site provides daily commentary continuously. On weekends the website focuses more on interacting with readers, with Tanenbaum and Bates answering reader questions on Saturdays and posting a mailbag of reader letters on Sundays.
Several political commenters, such as Chris Weigant, have noted the helpfulness of the website in aggregating polls state by state in order to predict the outcomes of elections.
Staff
Through most of the 2004 campaign, Andrew Tanenbaum went by the alias "the Votemaster", keeping his identity a secret. Tanenbaum is a civil libertarian, a member of Democrats Abroad (he is a long-time resident of the Netherlands) and generally supports Democratic candidates for office. He revealed his identity on November 1, 2004, as well as stating his reasons and qualifications for running the website.
In 2015, Tanenbaum was joined by Christopher Bates, who contributes to the site under the name Zenger, chosen in honor of John Peter Zenger. Their respective contributions are differentiated by (V) and (Z).
History
2004 Presidential election
The site began operating on May 24, 2004 with a simple map and a few links to other pages. During the months leading up to the 2004 U.S. Presidential election, the site was updated daily to reflect new state polls. The site was the most popular election site in the country, in the top 1,000 Web sites in the world, and in the top 10 blogs in the world. In November 2004, the website had more than 650,000 hits per day.
The main page consisted of a map of the United States with the individual states colored varying shades of red or blue, based on the polls for that state. For instance, Illinois, a state that was polling strongly for Democrat John Kerry was colored dark blue, whereas Michigan where Kerry's lead polled by a small margin was colored light blue. Analogously, Texas was dark red during the whole campaign, indicating Bush's strong lead there. All of the polling data were provided in multiple formats, including HTML, Excel, and .csv for downloading. Other features included historical data on previous elections, charts and animations showing the polls over the course of time, cartograms, and links to hundreds of other pages and external Websites with tables, charts, graphs, and other election data and information.
The main algorithm just used the most recent poll(s) in every state. If two polls came out on the same day, they were averaged. This algorithm used all published polls, including those by partisan pollsters such as Strategic Vision (R) and Hart Research (D). A second algorithm used only nonpartisan polls and averaged all polls during the past three days. A third algorithm used historical data to predict how undecided voters would break. Maps for each of the algorithms were given every day, but the first one got most of the publicity since it was on the main page.
The site's final tracking using algorithm 1 posted on Election Day, November 2 gave 262 electoral votes to John Kerry and 261 to George W. Bush, with 15 tossups. The second algorithm (averaging 3 days worth of nonpartisan polls) gave Kerry 245 and Bush 278 with 15 tossups. The third algorithm (predicting the undecideds) predicted 281 for Kerry and 257 for Bush.
The actual vote gave Kerry 252 to Bush's 286. Using nonpartisan polls and averaging a few days worth of polls did best. This algorithm got 47 states plus D.C. right, 1 state (Iowa) wrong, and said New Mexico and Wisconsin were too close to call. The most-recent-poll wins algorithm got 46 right, 4 wrong, and one too close to call.