1. Since Nader’s third party run in the 2000 election, the spoiler effect has become an important campaign tactic among political consultants.

2. U.S. campaigns have been run by professional consultants since at least 1896, when William McKinley, with the aid of Mark Hanna, defeated William Jennings Bryan for president. [p.93]

3. But modern political consulting, based on electronic media, scientific polling, and technical (game-theoretic) strategizing, only developed as a profession in the 1960s.

4. The term “political consultant” was coined by Joseph Napolitan, who established the scientific principles of the discipline in his 1972 book The Election Game and How to Win It. [p.94]

5. In 2008, there were something like 7,000 political consultants in the U.S., numerous college degree programs in the field, a trade journal (Campaigns and Elections), and professional organizations.

6. The professionalization of the field radically changed the nature of political campaigns in the last 3rd of the 20th century. And for Poundstone, Lee Atwater more than anyone else is responsible for this.

1. Poundstone claims that Atwater combined science and sleaze [p.95] by using polls to find issues that voters disagreed with an opposition candidate on, and making the issues the theme of the campaign.

2. For Atwater this meant relentless negative campaigning. Negative campaigning has always existed, but in the middle of the 20th c. campaigning in the U.S. became more civil than before or since.

3. Poundstone thinks TV produced more civility. [p.96] Before TV, people spent a lot of time in public places arguing about politics and rumor and gossip played important roles in opinion-making.

4. The rise of mass media, air-conditioning, the move to the suburbs, and automobile culture, produced a less public, more private culture in which politics was less important than before.

5. The first modern consultants sold candidates on TV. Broadcast laws required equal time for opposing views. In the 1980s, Reagan ended equal time laws, expanding the range of political commentary.

6. Relaxed equal-time rules made Fox News and Rush Limbaugh possible. And cable TV and its profusion of channels, and the internet, with its profusion of web sites, made rumor and gossip influential again.

7. In 1978, Lee Atwater began negative campaigning based on polls by appealing to people’s religious intolerance. [p.97] After this, he usually exploited basic types of intolerance among the voters.

8. [p.98] Atwater invented push-polling, claims Poundstone, where you pose as an independent pollster and ask confusing questions designed to leave a negative impression about your political opponent.

9. Atwater first used push-polling in 1980 for the campaign for South Carolina congressman Floyd Spence. Atwater help create the Southern Strategy of the GOP to turn the South Republican. [p.99]

10. In the 1980s, Atwater was a campaign consultant for Ronald Reagan and George Bush Sr. George Bush Jr. worked with Atwater and learned his campaign methods.

11. Atwater is most famous for his management of George Bush Sr.’s campaign in the 1988 presidential election. It set new standards for negative campaigning. [p.100]

12. Atwater first assembled a large group of more than 100 researchers to look for negative information about Bush’s Democratic opponent in the campaign – Massachusetts governor Michael Dukakis.

13. Atwater then tested all the negative facts they had found about Dukakis on a test group of Democrats who had voted for Reagan, but were leaning towards Dukakis in 1988. [p.101]

14. As governor, Dukakis had given convicted murderers weekend passes including one to William Horton, a black man who then killed a white man and raped his white wife. [p.102]

15. Atwater based the whole 1988 Bush campaign on this fact. Dukakis tried to ignore the negative campaigning and did not fight dirty himself. This was a big mistake. Bush won by a large margin. [p.103]

16. After the election, Atwater’s “Willie” Horton ads became the model for negative campaigning. And fighting dirty in political campaigns became the rule in politics after this. [p.104]

17. Atwater became head of the Republican National Committee in 1989. He tried to recruit Blacks to the party by kicking David Duke out of the party; this just made Duke’s supporters more likely to vote.

18. [p.105] The lesson of the 1988 election is that you had to go negative, if only to defend yourself, or lose. In 1976, 35 % of political ads were negative. In 1988, 83% were. By 2006, it was more than 90%.

19. [p.106] After Atwater and 1988, campaigning became a game where consultants had to do anything that would give them an advantage. Other than that, there were no rules at all.


Karl Rove and Lee Atwater were about the same age, but Atwater was internet and Rove worked in direct mail. [p.108] Direct mail continued to be important in the internet age. [p.108]

2. Rove took an analytic approach of who responds to what pitch with how much money.

3. Rove made the 2000 race the first to scientific test of the effectiveness of campaign techniques by setting aside some voter lists as controls and trying campaign techniques on others to see what worked and what didn’t. These tests were called “metrics.”

4. One finding is that only 6% of the voters were true swing voters. Rove summed this up by saying “There is no middle.”

5. Analysts studied the 2000 Nader spoiler effect where the Green Party had effectively elected Bush. And beginning in 2002, Republicans started aiding potential spoilers who might hurt Democrats.

5. The trick of using spoilers and splitting the vote has a long history. This included the 1908 presidential election between Taft and William Jennings Bryan, where [p.109] Socialist Eugene Debs toured the country in a train car, giving speeches. [p.110]

6. The Debs vote tipped a few extra states into Taft’s column but was not decisive. Taft won overwhelmingly.

7. Modern attempts to game spoilers date from 2002 when the New Mexico Republican party offered the Green Party money to run candidates in New Mexico’s 1st and 2nd Congressional Districts.

8. The green’s rejected the offer, and the story made it to the press. Karl Rove seems to have been the person who suggested making the offer.

In 1991, Rove had backed 5 Republicans to run in the Texas democratic primary for agricultural commissioner, acting as spoilers to unseat Democrat Jim Hightower from the job. [p.111]

1. The Green Party did not nominate Nader to run for president in 2004. But Nader ran as an independent. Without the backing of the Green party, Nader had less support – fewer people would sign petitions to put him on the ballot as a candidate or vote for him and he had less money.

2. Nader had to hire Republican consulting firms to collect signatures on ballots in Arizona as no Democratic firm would work for him. [p.112] The firms were JSM in Florida and Sproul.

3. JSM circulated Nader petitions at the same time they collected signatures for an anti-immigration ballot measure – many signed both. [p.113] Sproul was paid by the Republican National Committee.

4. What happened in Arizona for Nader was repeated across the nation in a Republican effort to get Nader on the ballot. [p.114] This was accompanied by a Democratic effort to keep Nader off the ballot.

5. In Oregon, candidates had to assemble 1000 registered voters in a nominating convention and have them sign a petition. Nader did this easily in 2000, but failed in April 2004. He tried again in June. [p.115]

6. Even with Republican support, Nader failed to make the ballot in June. Partly, this was due to Nader’s people not turning in the signed petitions, perhaps because many signers were registered Republicans.

7. In Michigan, Nader petitions were available at Republican Victory Centers. The Michigan GOP collected enough signatures to put Nader on the ballot. [p.116]

8. At first, Nader’s Michigan group refused to take the Republican signatures, but in the end relented.

9. Even with Republican help, Nader still failed to make the ballot in Arizona and Oregon, though he did, again with Republican help, make it in Michigan and New Mexico.

1. Ray Moore was another potential spoiler in the 2004 election. [p.117]

2. Moore was a judge for the state of Alabama who put a plaque of the 10 commandments up in his courtroom. The ACLU sued, a federal judge ordered it removed, Moore refused to remove it.

3. This made Moore famous. He ran for Alabama Supreme Court chief justice, beating Rove’s candidate, and then had a 2-ton version of the plaque put up in the Alabama Judiciary Building rotunda.

4. In 2004, Moore considered a run for president on the Constitution Party ticket. The Constitution Party was the 3rd largest party in the U.S. With 320,000 members, they were larger than the Green Party.

5. [p.118] A run by Moore threatened to split Rove’s conservative coalition between social and economic conservatives. The Constitution Party and Moore would draw off social conservative votes.

6. Rove planned to turn out the voters Moore appealed to; he couldn’t attack Moore without losing those voters. Bush was courting them by supporting a constitutional amendment to ban gay marriage.

7. [p.199] Democrats urged the party to support Moore as a potential spoiler to get Kerry elected. Moore decided not to run, and the Constitution Party candidate received only .12% of the vote.

8. Bush was popular in 2004 and did not need Nader to win. His popular vote margin over Kerry was 7x the Nader vote and the Libertarians, who drew off Bush voters, received almost as many votes as Nader.

9. Nevertheless, the Republicans’ furtive attempt to aid Nader in 2004 established the practice of aiding spoilers as a standard campaign strategy after that, and it has been practiced openly since that time.


1. [p.120] A special election is called for June 2006 in California’s 50th Congressional district to fill the seat. It is a prosperous Republican district. The Republican candidate is former congressman Brian Bilbray.

But it is a close race between Bilbray and his Democratic opponent Francie Busby. And it is a big race: The GOP raises 5 million for Bilbray, and Busby raises 2 million. [p.121]

2. Bilbray is a moderate Republican in general but an immigration hawk.

3. There is a potential spoiler in the election – William Griffith, a Republican running as an independent on an anti-immigration platform. He claimed to be more anti-immigration that Bilbray.

4. Griffith had 4% of the vote and had spent $2,000 of his own money on gas and a website. Voters in the 50th District began getting phone calls urging them to vote for Griffith, paid for by Busby. [p.122]

5. The point was to get enough voters to vote for Griffith so that Busby would win.

1. In a Nov 2005 Virginia election for Governor, voters received an “Official Democratic and Progressive Voter Guide” that criticized Democratic candidate Tim Kaine and supported Independent Randy Potts.

2. Potts was a liberal Republican running as an Independent. He had not sent voters the guide. They had been sent by Republican nominee Jerry Kilgore, who was nearly tied with Kaine in the polls.

3. Kilgore was hoping to split the Kaine vote. [p.123]

1. One attorney blogger called 2006 “The Year of the Spoiler.”

1. In a Pennsylvania 2006 election for Senator, Republican incumbent Rick Santorum was trailing Democratic challenger, who was the son of a former PA governor, Bob Casey, Jr.

2. Both candidates were pro-life Roman Catholic, pro-Iraq war anti-gun control. [p.124] The only liberal in the race was Green Party candidate Carl Romanelli.

3. The Green Party had little chance of getting the needed number of signatures to get on the ballot (67,070), until conservative Republican donors contributed $66,000 to Romanelli’s petition drive.

4. The money was given to JSM, a company that had collected signatures for Nader in 2004. JSM turned in 93,000 signatures for Romanelli. [p.125]

1. The new trend is about money. $1,000,000 of campaign money might get one an added 1% of the vote. But $66,000 to a 3rd party candidate might take 5% of the vote away from an opponent.

2. Gaming the spoiler effect is more cost-effective than conventional campaign techniques. For consultants, this is the real lesson of spoiler elections.

1. There were 5 candidates in the 2006 primary election for Arizona’s 8th Congressional District. The most popular candidate was the most conservative and the most unelectable. [p.126]

2. Polls found that Randy Graf, founding member of the Minuteman boarder vigilantes, was most likely to win the primary and most likely to lose to Democratic candidate Gabrielle “Gabby” Giffords.

3. The National Republican Congressional Committee tried to get rid of Graf by backing a moderate Republican, Steve Huffman, who was the only Republican with a chance of beating Giffords.

4. Huffman also would have likely beaten Graf too, but his vote was split by a second moderate Republican, Mike Hellon. The NRCC put pressure on Hellon to drop out but he ignored it.

5. before the primary, the DNC (Democratic National Committee) had paid for attack ads against Huffman to help Graf (who the Democratic candidate could easily beat).

6. With money and help from the DNC, Graf won the election against Huffman. [p.127]

1. In the 2006 Texas election for Governor, incumbent Rick Perry faced Democratic candidate Chris Bell, and two strong independents who also running, Carole Strayhorn and Kinky Friedman.

2. Strayhorn, former Austin mayor and state comptroller, had planned to challenge Perry in the Republican primary, but he was likely to win, so she dropped out to run as an independent.

3. Kinky Friedman was a colorful country rock musician who appealed to citizens who usually didn’t vote. Each candidate collected enough signatures to make the ballot.

4. In January 2006, The Dallas Morning News analyzed Strayhorn’s contributions [p.128] and determined that more than half her funds came from Democrats.

5. Polls found that Perry was first, Bell second, and Strayhorn 3rd or 4th. There was also a Libertarian Party candidate threatening to be a spoiler. Democrats were just hoping to shake things up for Perry.

1. In the 2006 Oregon race for Governor, independent Ben Westlund was polling 10% in a tight race between Democratic incumbent Ted Kulongoski and Republican challenger Ron Saxton.

2. Westlund thus dropped out to keep from being a spoiler. [p.129] That left two other potential spoilers in the race: the Constitution Party’s Mary Starrett.

3. Starrett seemed to be throwing the race to Kulongoski. She was also heavily funded – by the Democrats.

1. Poundstone said that 2006 is the year that the spoiler effect went mainstream for both parties “as a tool for political strategizing.” [p.130]

2. It is hard to imagine a law against it working. People can donate money to whomever they want.

3. Is there another way of voting that will be fairer and eliminate the spoiler effect?


1. [p.134] Condorcet was an 18th c. French mathematician (1743 – 1794). He was a classical liberal – he believed in individual liberty for everyone (rich, poor, men, women, white, black etc). [p.135]

2. He wrote the bill or rights for the French revolutionary constitution. But he did not believe in religious tolerance. (For the English, liberty was first of all about religious freedom.) [p. 136]

3. Jean-Charles de Borda (1733-1799) was also a French mathematician. He, along with Lavoisier and Laplace, is responsible for creating the metric system of measurement. [p.137]

4. Borda wanted to create a scientific system of voting. In a Royal Academy lecture given Jun 16 1770, he argued that democratic voting, as it was commonly understood and practiced, was unfair. [p.138]

5. In the talk, Borda proposed what he thought was a better voting method. He gave the talk at the Royal Academy again in 1784. It was then published in the Academy’s journal, Memoires de l’Académie.

6. Borda argued that while a plurality election is fair when then there are two candidates, it can lead to error in other cases, due to vote splitting. Then, the less popular person can sometimes win.

7. Vote splitting occurs when 2 candidates appeal to the same constituency thus splitting the vote of that constituency, leading to the election of a less popular third candidate. [p.139]

8. Vote splitting is found to regularly occur in Academy Award elections, thus leading to the Oscar being given to a less-highly regarded film than the films that split the vote. [p.140]

9. Borda proposed a solution to this problem, called “the Borda count,” where voters rank all the candidates – 1 for their 1st choice, 2 for the 2nd choice, etc. The candidate with the lowest score wins.

10. When there are more than 2 candidates, voters can express their preferences more completely than with a plurality vote. Then, a candidate most disliked will probably lose even if the other 2 split the vote.

11. In the case of David Duke running against Edwards and Roemer, if the majority hated Duke, but Edwards and Roemer had split the vote, Duke could have won in a regular plurality voting system.

12. With Borda’s method Duke would lose. [p.141] But Borda’s system of voting has several flaws.

13. Borda published his essay on his system of voting in 1784. The next year (in 1785) Condorcet published his own system of voting in a book about applying probability theory to society.

14. Condorcet’s system of voting was also meant to solve the problem of vote splitting. [p.142]

15. Like Borda, Condorcet said there was no problem with plurality voting when there were only two candidates. But when there are more than two candidates, the spoiler effect soon arises.

16. To solve the spoiler problem in elections with 3 or more candidates, Condorcet proposed having 2-way votes between all the candidates. The winner is the one who beats the others in 2-way votes.

17. For example, in the 1991 Louisiana governor’s race, Buddy Roemer probably would have beaten Edwards in a 2-way vote, and almost certainly would have beaten Duke in a 2-way vote.

18. So in applying the Condorcet voting system to Louisiana, Roemer, probably the candidate preferred by the most people, would have won.

19. Note: the Borda ranked voting system can be used to determine the Condorcet winner too, that is, it can be used to determine which candidate would win when every 2-way vote is counted.

20. One problem with these voting systems is that at the time it was difficult to tally Borda votes, and nearly impossible to tally Condorcet votes. With computers, these are not such great problems. [p.143]

21. Which system is fairer? They both claim to pick to most popular candidate in a 3-way election. But in some cases, they give different answers.

22. Condorcet showed that in some 3-way elections, candidate A will beat candidate B and in fact will win all the 2-way votes, yet with the Borda system, candidate B will still win.

23. Here is the example that William Poundstone, the author of Gaming the Vote, uses in his book:

In an election between Adams, Bush, and Clinton (A, B, and C), the results are:

Adams > Bush > Clinton – 30 votes
Adams > Clinton > Bush – 1 vote
Bush > Adams > Clinton – 29 votes
Bush > Clinton > Adams – 10 votes
Clinton > Adams > Bush – 10 votes
Clinton > Bush > Adams – 1 vote

In a Borda count, the winner gets 2 pts, 2nd place gets 1 pt, and the 3rd place none. That comes to 109 for Bush, 101 for Adams, and 33 for Clinton, so by Borda’s system, Bush wins. But for Condorcet, Adams wins 41 to 40 in 2-way votes. [p.144]

24. Worse, in a Borda count, the votes flip-flop if Clinton withdraws. Then, Adams wins 41-40. But the people’s preference for Adams versus Bush should not depend on whether or not Clinton runs. At least this doesn’t occur in the Condorcet system.

25. So we have a problem that is similar to vote splitting (though it is not clear how likely it is to occur or if it is as likely to occur as vote splitting does in a 3-way race).

26. What is special about the problem is that it occurs only when the voting is not transitive. Typically, where Bush is a conservative, Adams a liberal, and Clinton a super-liberal, those who prefer Clinton as a first choice, would prefer Adams over Bush as a second choice. That’s transitive voting; it’s our idea of rational.

27. But in non-transitive cases, some prefer Clinton to Adams and Adams to Bush, but Bush to Clinton. These cases are what create the problems for the Borda system and other voting systems.

28. Condorcet was the first to discover this problem with intransitive voting. It is called a “Condorcet cycle” and it is the same problem that the paradox of Arrow’s Theorem is based on.

29. The Condorcet cycle is not just a failure of Borda’s system – it is a problem for almost any voting system. It arises whenever a majority favors A over B and B over C yet favors C over A.

30. And when a Condorcet cycle occurs, no one is undefeated, so no one is the Condorcet winner.

31. However, it was not this problem, but another one that killed the Borda count. It was pointed out by the Marquis de Laplace that the Borda system is easily manipulated. [p. 145]

32. For example: in a tight race for NYC mayor between Dante De Blasio and Ivanka Trump, where the Borda system is being used, you rank all candidates, including 3rd-parties with no chance of winning.

33. Suppose in this election, David Duke is running for Mayor on the Nazi Party ticket. My preferences, then, are 1. Dante De Blasio, 2. Ivanka Trump, and 3. David Duke. Again, Duke has no chance of winning.

34. Now there’s a way to help Dante – that’s by ranking him 1st, Duke 2nd, and Ivanka 3rd. That lowers the Borda total for Ivanka without giving it to someone who can win, for Duke has no chance of winning.

35. But if everyone for Dante gives their 2nd vote to Duke and everyone for Ivanka gives their 2nd choice vote to Duke (so they don’t help Dante), and a few Nazi’s rank Duke # 1, Duke could win. [p.146]

36. Another alternative is if only some but not all voters “bury their votes” by voting for Duke as their 2nd choice, rather than voting for their true 2nd preference. Then Dante or Ivanka, but not Duke, wins.

37. The alternative sounds better, but if only some but not all voters “bury their vote” and Duke does not win, the likely winner is the one with the most dishonest voters. Manipulating the vote still pays off.

38. Furthermore, in most places where the Borda system is used, the voters almost always, immediately, intuitively start manipulating the vote to bury the threatening second choice. [p.147]

39. Both Borda and Condorcet voting systems had been used before Borda and Condorcet. The Roman Senate, for example, used the Borda system in the second century AD.

40 Furthermore, medieval writers Ramon Llull and Nicholas of Cusa wrote about such systems, and the Borda system was found to be in use on the South Pacific island of Kiribati, where the ruling party manipulated the vote in order to win in 1991. [p.148]

41. In effect, the system allows popular candidates to be eliminated by backroom political maneuvering (secret decisions for a group to bury their votes by giving their competitor a zero for last choice).

42. Jefferson and Madison were aware of the French controversy over voting systems. Both Borda’s and Condorcet’s methods became obsolete and were, for the most part, forgotten.

[JO. I think that burying votes in the Condorcet system similarly makes it possible for a Nazi 3rd party candidate to win.]