Annals of
Improbable Research Online (October 20, 2003)
An Algorithm for
Determining the Winners of U.S. Presidential Elections
Daniel Debowy
We present an algorithm for determining the winners of United States presidential elections, based on the previous experience of the major party candidates for President and Vice President. The algorithm correctly determines the winner of each of the 54 U.S. presidential elections between 1789 and 2000. Our algorithm predicts that President George W. Bush and Vice President Richard B. Cheney will win the 2004 election unless:
the Democratic nominee for President is Howard B. Dean,
the Democratic nominee for President is Wesley K. Clark and the Democratic nominee for Vice President has been Vice President for at least two years, a governor for at least five years, or a U.S. Representative for at least five years,
the Democratic nominee for President is Richard A. Gephardt and the Democratic nominee for Vice President is a banker, a college or university chancellor or president, or the child of a U.S. Senator, or
the Democratic nominee for Vice President is Albert A. Gore, Jr. or John D. Rockefeller, IV, and the Democratic nominee for President has not been divorced, has not been a special prosecutor, and is a Protestant, Deist, or Catholic.
Although any of the currently declared Democratic candidates for President could, in theory, win in 2004 if they carefully choose their vice presidential candidates, in practice it would be difficult for many of them to find candidates for Vice President with the right combination of governmental and non-governmental experience.
1.
Introduction
Throughout the history of the United States, politicians have attempted to determine the likelihood of one candidate or another winning a presidential election. Schulman (2001) presented an algorithm that purported to do just that for the presidential elections since 1932. Their algorithm depended on the number of years the candidates had served as President, Vice President, U. S. Senator, U. S. Representative, and Governor. It also took into account whether the candidate had been Director of Central Intelligence, a general officer in the United States Armed Forces, and/or had ordered the combat use of nuclear weapons. Although their formula correctly predicted the winners of the U.S. presidential elections between 1932 and 2000, it did not correctly predict the winners of all the U.S. presidential elections between 1789 and 1928. Schulman (2001) obviously believed that U.S. presidential elections over the last 70 years are not typical of all U.S. presidential elections. We disagree with that conclusion, and in this paper present an algorithm that correctly determines the winners of every U.S. presidential election between 1789 and 2000.
2.
Methods and Results
We analyzed the experience of the major party candidates for President and Vice President in each of the U. S. Presidential elections since 1804, and the experience of the two top candidates for President in 1789, 1792, 1796, and 1800 (when the candidate who received the most electoral votes became President, and the candidate who received the next largest share of electoral votes became Vice President). We discovered the following empirical formula after an extensive phase space search:
Presidential Electability = 5*(years as President) + years as U.S. Representative + 11*(years as Governor),
+110 if the candidate has been a four- or five-star general officer in the United States Armed Forces,
+110 if the candidate has been a college or university president or chancellor,
+110 if the candidate is the child of a U.S. Senator,
110 if the candidate has been divorced,
110 if the candidate has been a special prosecutor,
110 if the candidate was the first adherent of a particular religion (e.g., Protestantism, Deism, or Catholicism) to be a major-party candidate for President,
110 if the candidate was an officer of a lobbying organization at the time of the election.
Vice Presidential Electability = 4*(years as Vice President) + years as U.S. Representative + years as Governor,
+110 if the candidate has been a corporate banker,
+110 if the candidate has been a college or university president or chancellor,
+110 if the candidate is the child of a U.S. Senator,
110 if the candidate was the first adherent of a particular religion (e.g., Protestantism, Deism, Catholicism, or Judaism) to be a major-party candidate for Vice President,
110 if the candidate was an officer of a lobbying organization at the time of the election.
Total Electability = Presidential Electability + Vice Presidential Electability.
Years in office is equal to the number of years the candidate served in a particular office, rounded up as long as the partial year service was one month or more, unless the candidate moved directly from one public office to another, in which case the office in which the candidate spent a larger fraction of their time during that year receives credit for the year. Years of service for federal offices were verified using the Biographical Directory of the United States Congress and years of service for governor were verified using the relevant state websites. In each U. S. presidential election between 1789 and 2000, the candidates with the higher total electability won, as seen in Table 1 below. Note that the electorate doesnt appear to care one way or the other how long a candidate has served as a U.S. Senator, but we included that information in Table 1 for completeness.
|
Year |
Candidate |
Pres. |
V.P. |
Sen. |
Rep. |
Gov. |
Other |
Total |
||
|
1789 |
George Washington |
0 |
0 |
0 |
0 |
0 |
College Chancellor1 |
110 |
110 |
|
|
John Adams |
0 |
0 |
0 |
0 |
0 |
|
0 |
0 |
||
|
1792 |
George Washington |
4 |
0 |
0 |
0 |
0 |
College Chancellor1 |
110 |
130 |
|
|
John Adams |
0 |
4 |
0 |
0 |
0 |
|
0 |
0 |
||
|
1796 |
John Adams |
0 |
8 |
0 |
0 |
0 |
|
0 |
0 |
|
|
Thomas Jefferson |
0 |
0 |
0 |
0 |
2 |
First Deist |
-110 |
-88 |
||
|
1800 |
Thomas Jefferson |
0 |
4 |
0 |
0 |
2 |
|
0 |
22 |
|
|
John Adams |
4 |
8 |
0 |
0 |
0 |
|
0 |
20 |
||
|
1804 |
Thomas Jefferson |
4 |
4 |
0 |
0 |
2 |
|
0 |
42 |
64 |
|
George Clinton |
0 |
0 |
0 |
0 |
22 |
|
0 |
22 |
||
|
Charles C. Pinckney |
0 |
0 |
0 |
0 |
0 |
|
0 |
0 |
0 |
|
|
Rufus King |
0 |
0 |
7 |
0 |
0 |
|
0 |
0 |
||
|
1808 |
James Madison |
0 |
0 |
0 |
8 |
0 |
|
0 |
8 |
46 |
|
George Clinton |
0 |
4 |
0 |
0 |
22 |
|
0 |
38 |
||
|
Charles C. Pinckney |
0 |
0 |
0 |
0 |
0 |
|
0 |
0 |
0 |
|
|
Rufus King |
0 |
0 |
7 |
0 |
0 |
|
0 |
0 |
||
|
1812 |
James Madison |
4 |
0 |
0 |
8 |
0 |
|
0 |
28 |
34 |
|
Elbridge Gerry |
0 |
0 |
0 |
4 |
2 |
|
0 |
6 |
||
|
DeWitt Clinton |
0 |
0 |
2 |
0 |
0 |
|
0 |
0 |
0 |
|
|
Jared Ingersoll |
0 |
0 |
0 |
0 |
0 |
|
0 |
0 |
||
|
1816 |
James Monroe |
0 |
0 |
4 |
0 |
4 |
|
0 |
44 |
54 |
|
Daniel D. Tompkins |
0 |
0 |
0 |
0 |
10 |
|
0 |
10 |
||
|
Rufus King |
0 |
0 |
11 |
0 |
0 |
|
0 |
0 |
3 |
|
|
John E. Howard |
0 |
0 |
7 |
0 |
3 |
|
0 |
3 |
||
|
1820 |
James Monroe |
4 |
0 |
4 |
0 |
4 |
|
0 |
64 |
90 |
|
Daniel D. Tompkins |
0 |
4 |
0 |
0 |
10 |
|
0 |
26 |
||
|
John Q. Adams |
0 |
0 |
6 |
0 |
0 |
|
0 |
0 |
2 |
|
|
Richard Stockton |
0 |
0 |
3 |
2 |
0 |
|
0 |
2 |
||
|
1824 |
John Q. Adams |
0 |
0 |
6 |
0 |
0 |
|
0 |
0 |
7 |
|
John C. Calhoun |
0 |
0 |
0 |
7 |
0 |
|
0 |
7 |
||
|
Andrew Jackson |
0 |
0 |
3 | |||||||