Thomas Dewey was relentlessly ambitious. He was always in a hurry, always striving for success, always pushing to reach the top.
His tiny hometown of Owosso, Michigan, offered no opportunities for fame or fortune, so Dewey sought a larger stage after graduating from college in 1923. It came as no surprise to family and friends that he opted for the largest venue of all, New York City, which he quickly pronounced “a thousand times better” than Owosso.1 Dewey was not yet certain of his course—he might employ his rich baritone voice as an opera singer or his imposing intellectual skills as a lawyer—but he was absolutely determined to strike it big.
Fate made the choice for him. Incipient laryngitis struck on the morning of his first solo recital in New York, inspiring serious doubts about the riskiness of an operatic career. The law seemed a much safer option, and Dewey proved to be remarkably good at it. His organizational ability, commanding presence, and fearless tenacity would serve him well in the court-room. He had found his path.
But nobody, not even Tom Dewey, could have anticipated a rise so stunningly meteoric. He exploded into national prominence while still in his 30s, a brisk, no-nonsense prosecutor who grappled with the toughest figures in the underworld—mobsters like Lucky Luciano and Dutch Schultz—and sent most of them to prison. The young district attorney became known as the “gangbuster,” the preeminent media idol of his day. “If you don’t think Dewey is Public Hero Number One,” the Philadelphia Inquirer wrote, “listen to the applause he gets every time he is shown in a newsreel.”2
The Republican Party sought to capitalize on his massive popularity, nominating Dewey for governor of New York in 1938 at the politically tender age of 36. The Democratic incumbent, Herbert Lehman, was believed to be invincible, having trounced his previous opponent by 520,000 votes. Dewey whittled the margin to 65,000, coming within one-and-a-half percentage points of victory. His performance was so impressive, even in defeat, that it buoyed spirits at Republican headquarters. “You’re a sure bet for 1940,” someone shouted as Dewey walked to the podium to concede defeat.3 The next race for governor wasn’t slated until 1942. The enthusiastic supporter clearly had a different—and bigger—target in mind.
As did Dewey, whose ambition knew no bounds. He pursued the presidency in 1940, even though two unresolved questions cast a shadow over his prospects. He wasn’t a governor or a senator, merely the district attorney of Manhattan. Would people be willing to vote for a DA for president? And he would celebrate his 38th birthday just three months before the convention. Would voters accept a presidential candidate so young? Harold Ickes, the acerbic secretary of the interior in Franklin Roosevelt’s cabinet, was dubious on both counts. He joked that Dewey had “thrown his diaper into the ring.”4
The campaign went better than the skeptics expected. Dewey emerged as the Republican front-runner, dropping back to the pack only after the Nazis invaded France in May 1940. Ickes proved to be correct in the end. Americans didn’t believe a candidate still in his 30s could be entrusted with a challenge as monumental as World War II. The older, more experienced Wendell Willkie won the nomination at the Republican convention, then lost the general election to the ultimate father figure, Roosevelt, who had guided the nation through the turbulent 1930s.
Dewey viewed his defeat as a temporary setback. He redoubled his efforts, first winning the governorship of New York in 1942, then securing the next two Republican presidential nominations. The odds were against him in the general election of 1944. Roosevelt was again the Democratic nominee, American forces were still battling their way toward Berlin and Tokyo, and the nation was reluctant to change horses in midstream. But 1948 was different. The war was over. The new president, Harry Truman, was decidedly unpopular. Voters seemed anxious for an alternative. Newsweek polled 50 political experts a month before the election. All 50 said Dewey would win.5
Yet he lost.
Dewey affected a public attitude of nonchalance, but it was a stunning, demoralizing, incomprehensible defeat. He was suddenly a political has-been at age 46. The presidency would never be his. Hindsight revealed that Truman’s aggressive jabs had scored with the voters, while Dewey had frittered away the lead by playing it safe. The Republicans, according to Truman, had “stuck a pitchfork in the farmer’s back”6 and had forced consumers to swallow “the economic tapeworm of big business,” all to benefit the “bloodsuckers with offices in Wall Street.”7 Dewey, lulled by inaccurately favorable polls, had responded with soporifics. “I will not get down into the gutter with that fellow,”8 he privately told his aides.
Tom Dewey would live until 1971, giving him nearly a quarter century to ponder what might have been. He acknowledged his shortcomings as a campaigner, but also came to regret his precocity—“everything came too early for me”9—and to lose whatever patience he had once possessed for politics. He would have “to be touched with madness” to run for president again, he said in the 1950s. He recalled the whole experience with distaste—the speeches, the handshaking, the fund-raising, “being dragged from pillar to post, with no time to think,” and the vague sense that he was merely a player in a quadrennial diversion for the masses.10
“I have learned from bitter experience,” he said, “that Americans somehow regard a political campaign as a sporting event.”
Dewey had a point. There are many compelling similarities between politics and sports:
This final similarity, to be fair, deserves an asterisk. A comparable level of fascination does not guarantee a corresponding level of quality. Several sports, especially baseball, have set a brisk statistical pace that the political world has been unable to emulate. “Unlike baseball, where statisticians pore over everything from earned run average (traditional metrics) to batting average on balls in play (sabermetric bliss), politics has seen limited breakthroughs from obsessive data analysis,” concedes political reporter Sam Stein.19
This disparity is punctuated by one of Stein’s parenthetical adjectives. Statistical analysis has become so central to the study and enjoyment of baseball that it is now recognized as a distinct discipline, known as sabermetrics. The name was coined by Bill James, a night watchman at a Kansas pork-and-beans cannery, who began playing with numbers in an effort to unlock baseball’s mysteries.20 His first book of stats and observations, self-published in 1977, sold all of 75 copies. But word spread quickly, and James ascended to iconic status. Time magazine included him on its 2006 list of 100 people “whose power, talent, or moral example is transforming our world.”21
A flotilla of analysts now sails in James’s wake, churning out books, magazines, and websites devoted to sabermetrics. Baseball has always been a game of simple numbers—batting averages, win-loss records, and the like—but sabermetricians have devised dozens of specialized, highly complex statistics. Fans now toss around such acronyms as BABIP, FIP, OPS, and WAR when debating the relative merits of their favorite pitchers and hitters. WAR, which stands for wins above replacement, reduces the overall performance of any player to a single number. A benchwarmer with a WAR of 0.0 brings no value to his team, while the few men above 5.0 are all-stars. Babe Ruth topped 10.0 in nine different seasons, a true indication of dominance.
Nothing in politics can compare. “There is a whole industry in baseball dedicated to the proper understanding and interpretation of statistics,” Silver wrote in 2008. “In polling and politics, there is nearly as much data as there is for first basemen. In this year’s Democratic primaries, there were statistics for every gender, race, age, occupation, and geography—reasons why (Hillary) Clinton won older women, or Obama took college students. But the understanding has lagged behind.”22
Obama’s reelection effort did make strides in 2012. “We are going to measure every single thing in this campaign,” promised Jim Messina, the president’s campaign manager.23 His number crunchers used “big data” to identify potential Obama supporters, donors, and voters. They even developed a mathematical algorithm to determine the perfect hostess for a fund-raising event targeted at East Coast women in their 40s. The computer chose Sarah Jessica Parker.
But Obama’s data analysis differed from sabermetrics in two very important ways. The president’s team conducted its work in secret, and it focused solely on the election of 2012. Bill James and his acolytes publish their findings for everybody to see, and they embrace all 150 years of baseball’s history.
If you visit your local library or the Internet, you’ll notice the same disparity. Political books and websites don’t offer much to readers who are numerically inclined. Statistics are typically confined to state-by-state and county-by-county election returns. No specialized stats of any kind. No rankings of the best and worst performances of all time. No year-by-year records of presidential candidates. None of the originality and variety that sports fans take for granted.
This book is a humble effort to begin filling that void. Consider it an early step toward a political version of sabermetrics—raising new questions and offering tentative answers, while forging a numerical link between the past, present, and future. Some of the statistics contained within these covers might prove to be useful, but more importantly, I hope they’re intriguing, provocative, and fun. Perhaps they’ll inspire other fans of presidential politics to develop more and better metrics in the years ahead.
I have been on both sides of the politics/sports divide. My first book, The Pursuit of the White House, was a guide to the candidates who took part in the 50 presidential elections from 1789 to 1984. It received a smattering of positive notices, which I naturally enjoyed, though I most vividly recall a reviewer who was unable to suppress his ennui. “There is nothing seriously wrong with this book,” he sniffed. (High praise, indeed.) Subsequent books of mine have focused on the political ramifications of suburban growth, the lessons to be drawn from the life stories of American presidents, and the twists and turns that marked the famous electoral battle between John Kennedy and Richard Nixon in 1960.
But the best preparation for this book may have been Leveling the Field, my sole venture into sportswriting. I devised a system that equalized statistics throughout baseball’s history. Home runs, for instance, were remarkably rare during the “dead ball” era at the dawn of the 20th century. Tommy Leach led the National League with the less-than-imposing total of six homers in 1902.24 But the steroid era produced dramatically bigger numbers by century’s end, notably Mark McGwire’s 70 home runs in 1998. My calculations determined that Leach’s paltry total was actually the equivalent of 37 homers in McGwire’s time—not enough to lead the league, but certainly sufficient to impress the fans and intimidate most pitchers.
You’ll find a similar approach in this book. Vote totals will be equalized to allow simple and fair comparisons from election to election. Several new statistics—the political counterparts of WAR—will be introduced. Everything will be easy to understand. “As much as it might seem, sabermetrics is not rocket science,” says baseball analyst Mitchel Lichtman. “In the world of science and applied math, it is frighteningly simple, comparatively.”25 The same holds true for the metrics I have created.
You should be aware, however, that this is not a conventional history book. If you’re seeking the inside story of the 1892 Republican convention or a step-by-step analysis of the career of Alfred Landon, you have come to the wrong place. (You are to be congratulated, however, for choosing two topics guaranteed to induce a satisfying night’s sleep.) This is a compendium of original statistics, fresh interpretations of old numbers, and purely random observations. Consider it a new toolbox, chock-full of screwdrivers and wrenches that have been specially crafted for statistical tinkering.
These tools, if we use them with any skill, should provide us with new insights into the political process, as well as a better understanding of the men and women who have sought the presidency in the past (or will be seeking it in the near future). Tom Dewey, for instance, might have been less surprised and less embittered by his fate if he had been aware of the relevant numbers:
The remainder of this chapter will explain the logic behind PI and other statistical tools that have been devised to evaluate the vote-getting potential and electoral performance of presidential candidates. You’ll learn about PV*EQ and CS and CV% and ROP and several other indicators. You’ll also find out how to decipher the charts in the chapters that follow— and if there’s one thing this book has, it’s plenty of charts.
Let’s get started.
The Founding Fathers, despite their populist rhetoric, saw no reason to let common people participate in presidential elections. Most of the nation’s creators thought the very idea of popular voting was unnecessarily dangerous. “It would be [as] unnatural to refer the choice of a proper character for chief magistrate to the people,” Virginia’s George Mason scoffed, “as it would to refer a trial of colors to a blind man.”27
The Constitution circumvented this hazard by establishing the Electoral College, a small group of political and financial elites. Each of its members would cast two votes. The candidate receiving the most support would become president, while the runner-up would be vice president. This uncomplicated system worked to perfection when it debuted in 1789. All 69 electors voted for George Washington. John Adams, the new vice president, finished a distant second with 34 electoral votes. Nobody else received more than nine.
But the double-ballot system was inherently flawed, as it became clear in 1800. The Democratic-Republican Party’s respective nominees for president and vice president, Thomas Jefferson and Aaron Burr, received 73 electoral votes apiece, leaving it to the House of Representatives to break the tie. A week of legislative skirmishing ensued—36 roll calls in all—before Jefferson was declared the winner. The new president pushed Congress to enact the Twelfth Amendment prior to the election of 1804, henceforth requiring members of the Electoral College to cast separate votes for president and vice president.
This was the only significant alteration during the first three decades of presidential politics, though its impact was not especially large. The Twelfth Amendment, rather than greatly changing the original process, merely streamlined it. The elites, acting through the Electoral College, still held total control. The common people were still barred from participation.
That’s why the statistical remnants of the first nine elections (1789 through 1820) are so meager. These are the only numbers available for candidates in those races:
The array of political statistics would expand greatly in the 1820s and 1830s, thanks to a pair of new developments.
The everyday Americans who were helping to build the new country inevitably demanded more say in how it was run. Some states began allowing white men to vote for members of the Electoral College or even for presidential candidates themselves. This practice became sufficiently widespread by 1824 that meaningful national totals could be tabulated for the first time, thereby adding these two categories to the statistical lexicon:
The second innovation during that brief span had an unlikely genesis. The Anti-Masonic Party, a fringe movement energized by a rabid hatred of Freemasonry, decided to hold a national convention to choose its 1832 presidential nominee. The outcome should have raised doubts about the value of such gatherings. The Anti-Masons nominated William Wirt, a former Mason who fondly remembered his lodge as “a social and charitable club.”28 Yet the major parties were impressed. They had previously relied on informal networks of state leaders and congressional caucuses to anoint their presidential candidates. They now decided to convene national meetings of their own.
The first conventions of the Democratic and National Republican parties were harmonious. The Democrats unanimously nominated Andrew Jackson for a second term in 1832, and 167 of 168 National Republican delegates endorsed Henry Clay. But subsequent conventions were often drenched in acrimony or indecision. The Whigs, for instance, required 53 ballots to select a nominee in 1852, the same year the Democrats called the roll 49 times.
This book consequently offers two or four statistics for a given convention, depending on the number of ballots:
Convention delegates have always shown an unnatural fondness for casting fractional votes, but all ConFB and ConLB figures have been rounded to the nearest whole numbers. The official tabulation for the 1972 Democratic convention shows George McGovern with 1,728.35 votes, but I’ve settled for 1,728. (FB% and LB%, however, are based on the original, unrounded totals.)
The convention remained the sole vehicle for nominating presidential candidates into the 20th century. That began to change with the advent of the primary election in 1912. Primaries were relatively insignificant at first—serving as simple barometers of public opinion—but they grew to be much more important over time. No candidate can be nominated these days without running successfully in the primaries, especially the initial midwinter contests. The summer conventions merely rubber-stamp the decisions made by primary voters months before. There hasn’t been a multiballot convention since 1952, and it’s unlikely we’ll ever see one again.
The number of primary elections varies from year to year, based on the procedures followed by each state and each party. The Republicans conducted as many as 43 state primaries in 2000, but slipped back to 39 by 2012. The Democrats peaked at 40 in 2008, then dropped to 30 in the following election. These are the three relevant categories:
The Electoral College, popular voting, conventions, and primaries can be considered the four building blocks of presidential politics. The 12 basic statistics they’ve spawned—EV, PV, ConFB, PriV, and the rest—provide a quick snapshot of the relative success or failure of every candidate who has pursued the presidency since 1789, just as meat-and-potatoes figures like HR, RBI, and ERA have entertained baseball fans for generations.
More significant for our purposes, however, is the fact that these simple numbers can form the basis for statistics of greater complexity. We’ll turn to that subject once we address a question of vital importance: Which candidates should be included when we begin collecting data?
Registering as an official candidate for president is remarkably easy. Go to the Federal Election Commission’s website, download FEC Form 2, answer the eight questions, and mail the completed document to Washington. That’s all there is to it.
A total of 417 presidential hopefuls submitted Form 2 in 2012—spanning the alphabet from Dorothy Ann Adams and Yinka Adeshina to Michael Yost and William David Zollinger.29 Most of these self-proclaimed candidates went no further than filing their paperwork, though 44 did qualify for the Democratic or Republican ballot in New Hampshire, where the nation’s first primary election was conducted on January 10.
Forty-four is a sizable number, implying that New Hampshire’s voters enjoyed a broad array of viable presidential options. But that wasn’t the case, as anybody even remotely aware of political realities could attest. Media coverage focused on half a dozen candidates—Democrat Barack Obama and five Republican challengers—and ignored the other 38. The six front-runners consequently received 97 percent of all votes cast in the New Hampshire primary. The minor candidates averaged 278 votes apiece. Sixteen drew fewer than 100 votes.
New Hampshire and subsequent primaries eventually trimmed the race to a pair of major-party nominees: Obama and Republican Mitt Romney. But more than two dozen minor parties named their own candidates, dramatically expanding the field once again. Twenty-eight contenders appeared on the November 2012 ballot in at least one state.
The imbalance of support was even more pronounced in the general election than in the battles for the Democratic and Republican nominations. These were the final totals on November 6, 2012:
| Barack Obama (Democrat) | 65,915,796 | 51.06% |
| Mitt Romney (Republican) | 60,933,500 | 47.20% |
| Gary Johnson (Libertarian) | 1,275,971 | 0.99% |
| Jill Stein (Green, Pacific Green, Mountain) | 469,628 | 0.36% |
| Write-ins | 136,040 | 0.11% |
| Virgil Goode (Constitution, U.S. Taxpayers) | 122,388 | 0.09% |
| Roseanne Barr (Peace and Freedom) | 67,326 | 0.05% |
| Ross C. "Rocky" Anderson (Justice, Progressive) | 43,018 | 0.03% |
| Tom Hoefling (American Independent, America's Party) | 40,628 | 0.03% |
| Randall Terry (Independent) | 13,105 | 0.01% |
| Richard Duncan (Independent) | 12,557 | 0.01% |
| Peta Lindsay (Socialism and Liberation) | 7,791 | 0.01% |
| None of these candidates (Nevada) | 5,770 | 0.00% |
| Chuck Baldwin (Reform) | 5,017 | 0.00% |
| Will Christensen (Constitution) | 4,453 | 0.00% |
| Stewart Alexander (Socialist) | 4,405 | 0.00% |
| James Harris (Socialist Workers) | 4,117 | 0.00% |
| Thomas Robert Stevens (Objectivist) | 4,091 | 0.00% |
| Jim Carlson (Grassroots) | 3,149 | 0.00% |
| Jill Reed (Unaffiliated) | 2,875 | 0.00% |
| Merlin Miller (American Third Position) | 2,701 | 0.00% |
| Sheila "Samm" Tittle (We the People) | 2,572 | 0.00% |
| Gloria La Riva (Socialism and Liberation) | 1,608 | 0.00% |
| Jerry White (Socialist Equality) | 1,279 | 0.00% |
| Dean Morstad (Constitutional Government) | 1,094 | 0.00% |
| Jerry Litzel (Independent) | 1,027 | 0.00% |
| Barbara Dale Washer (Reform) | 1,016 | 0.00% |
| Jeff Boss (NSA Did 911) | 1,007 | 0.00% |
| Andre Barnett (Reform) | 956 | 0.00% |
| Jack Fellure (Prohibition) | 518 | 0.00% |
Obama and Romney combined for 126.85 million votes, accounting for 98.26 percent of all ballots cast in the 2012 election. The other 26 candidates—along with write-ins and Nevada’s “none of the above” option—divided the thin slice of pie that remained, a mere 1.74 percent. The average minor-party candidate received just 80,000 votes nationwide.
It’s quite clear that few aspirants are able to mount effective presidential campaigns. Most candidates are incapable of raising the necessary hundreds of millions of dollars, and they also lack the skill required to command the daily attention of major newspapers, wire services, and broadcast networks. For every Barack Obama or Mitt Romney who succeeds in political life, there are innumerable Barbara Dale Washers and Jack Fellures who fail to make any impression whatsoever.
That poses a serious problem for analysts of presidential elections. The glut of insignificant candidates greatly increases our workload, while burying important trends under a mountain of minutiae. If we were to break down the state-by-state results not just for Obama and Romney, but for all 28 general-election candidates, imagine how much time we would waste and how much useless data we would generate. It would be a classic case of not seeing the forest for the trees.
That’s why it’s essential to separate the serious contenders from the mass of minor-party fanatics, publicity seekers, and occasional lunatics who clog the presidential ballot. I’ve come up with five requirements—known as qualification levels—that address this need. If a candidate achieves any of these benchmarks in a given year, he or she is considered to be a qualified candidate:
A presidential hopeful needs to reach just one of these levels in a given year to be designated as a qualified candidate, thereby warranting serious and thorough consideration in this book. Anybody who falls short of all five levels is labeled as nonqualified and is essentially ignored. (You might be curious, by the way, about the precision in a few of these five rules. Why 1.95 or 3.95 percent? Because they round to 2 or 4 percent, that’s all.)
A total of 289 men and women have met these statistical standards in the 57 presidential elections since 1789, mounting 444 qualified candidacies in all. (Henry Clay and James Blaine share the record with five apiece.) Most of these contenders belonged to the seven major parties recognized by this book, which are abbreviated this way:
| D: | Democratic (1832–present) |
| DR: | Democratic-Republican (1789–1828) |
| F: | Federalist (1789–1816) |
| NR: | National Republican (1828–1832) |
| R: | Republican (1856–present) |
| SD: | Southern Democratic (1860) |
| W: | Whig (1836–1852) |
The only major party that might throw you for a loop is the Southern Democratic Party, which split off from the main Democratic organization in 1860, surely the most contentious year in American political history. The two factional nominees in that race—Democrat Stephen Douglas and Southern Democrat John Breckinridge—essentially battled to a tie. Douglas received more PVs, Breckinridge more EVs. They went head-to-head in 29 states, with Breckinridge drawing more votes in 15, Douglas in 14. It seems only fair to treat both as major-party candidates.
I should also explain why I date the Democratic-Republicans and Federalists back to 1789. No parties functioned as such during that first campaign, but not everything was sweetness and light. Opposing factions were already bickering, and much to George Washington’s great displeasure, they soon blossomed into full-blown parties. “How unfortunate and how much it is to be regretted,” he moaned.30 But this split was a fact of political life from the very beginning. I have acknowledged its existence by inserting the subsequent political affiliations of 1789’s qualified candidates.
The American political system has always had remarkable symmetry, starting with those early skirmishes between the Federalists and Democratic-Republicans. The former faded away after 1816, eventually replaced in turn by National Republicans, Whigs, and Republicans. The latter morphed into Democrats by 1832. This evolutionary process guaranteed that two major parties were (almost always) in the field at any given time, prepared to do battle.
This book recognizes 113 major-party nominees in the 57 presidential elections, an average of 1.98 per year. Nominations have been matters of official record since the convention made its debut in 1832, but a bit of subjectivity is needed to designate nominees for a few elections between 1789 and 1828 (and a couple that came later):
There are plenty of minor parties, too, as indicated by the official returns for the 2012 election. You’ll encounter 28 minor parties in the charts throughout this book. You’ll even find a few cases where a single candidate is tied to a major party and minor party in the same election. If the two parties are linked by a hyphen, the candidate ran on both tickets at the same time. If they’re separated by a slash, the candidate began in one party and switched in midstream to the second.
Here are the minor-party abbreviations:
| AI: | American Independent |
| AM: | Anti-Masonic |
| ASI: | America's Independent |
| AW: | American-Whig |
| CON: | Constitution |
| CSN: | Consumer |
| CU: | Constitutional Union |
| ER: | Economic Recovery |
| FL: | Farmer-Labor |
| FS: | Free Soil |
| GBK: | Greenback |
| GRN: | Green |
| I: | Independent |
| LIB: | Libertarian |
| LR: | Liberal Republican |
| LTY: | Liberty |
| NER: | National Economic Recovery |
| NU: | National Unity |
| P: | Progressive |
| POP: | Populist |
| PRO: | Prohibition |
| REF: | Reform |
| SOC: | Socialist |
| SR: | States' Rights |
| SRD: | States' Rights Democratic |
| TC: | Taxpayers and Constitution |
| U: | Union |
| USL: | U.S. Labor |
All of the groundwork has now been laid. We’ve browsed through the basic statistics of presidential politics, developed a system to determine which candidates deserve further study, and separated the major parties from the minor pretenders. It’s time to move on to more advanced concepts.
You could make a passable case that Barack Obama is America’s greatest vote-getter of all time. There have been only nine instances of a presidential candidate drawing more than 50 million popular votes in a single election, and Obama occupies the top two positions on the list:
| 1. | Barack Obama (D-2008) | 69,498,516 |
| 2. | Barack Obama (D-2012) | 65,915,796 |
| 3. | George W. Bush (R-2004) | 62,040,610 |
| 4. | Mitt Romney (R-2012) | 60,933,500 |
| 5. | John McCain (R-2008) | 59,948,323 |
| 6. | John Kerry (D-2004) | 59,028,439 |
| 7. | Ronald Reagan (R-1984) | 54,455,075 |
| 8. | Al Gore (D-2000) | 50,999,897 |
| 9. | George W. Bush (R-2000) | 50,456,002 |
What if we add up all of the popular votes every candidate received throughout his career? Obama heads the rankings again, holding a comfortably wide lead of more than 22 million votes:
| 1. | Barack Obama | 135,414,312 |
| 2. | Richard Nixon | 113,063,548 |
| 3. | George W. Bush] | 112,496,612 |
| 4. | Franklin Roosevelt | 103,425,434 |
| 5. | Ronald Reagan | 98,359,228 |
| 6. | Bill Clinton | 92,311,683 |
| 7. | George H. W. Bush | 87,989,988 |
| 8. | Jimmy Carter | 76,314,646 |
| 9. | Dwight Eisenhower | 69,368,417 |
| 10. | Mitt Romney | 60,933,500 |
The pro-Obama argument, of course, isn’t as strong as it first appears, which is why I used “passable” as an adjective in the lead sentence. Closer scrutiny proves that the single-year PV standings aren’t particularly compelling. Four of the nine men who topped 50 million votes actually lost their elections, hardly a sign of political supremacy.
The endless expansion of the American electorate favors recent candidates, which is why eight of the nine biggest PV totals have been recorded since 2000. A few decades can make an enormous difference in standards of excellence, as we see in a comparison of Lyndon Johnson and Obama. Johnson posted the highest PV% in history, an astounding 61.05 percent. But he did it in 1964, when just 70.6 million people voted:
61.05% • 70,644,592 = 43,129,566 votes
Obama fared about eight percentage points worse than Johnson, 52.93 percent in 2008, but the electorate had nearly doubled to 131.3 million voters:
52.93% • 131,313,820 = 69,498,516 votes
Johnson’s performance was obviously superior—vastly so—yet it’s Obama who holds the record for popular votes in a single election. You have to scan all the way down to 15th place to find LBJ, who is mired more than 26 million votes behind the leader.
The career standings are similarly flawed. Obama sits in first place simply because he went fishing in the largest pool of voters. The same principle has propelled several other Modern Era presidents into the career top 10—both Bushes, Clinton, Reagan, Carter, Nixon—while two-term giants from earlier eras, including Andrew Jackson, Abraham Lincoln, and Grover Cleveland, are nowhere to be seen. The strangest example of this chronological bias is Mitt Romney, who ranks 10th in career popular votes, even though he ran in a single general election, which he lost.
It’s clear that we need to find a better way to compare statistics from different political eras. We need a fair and accurate system that will level the field.
The federal government—yes, those faceless bureaucrats in Washington— can point us in the right direction. The Bureau of Labor Statistics (BLS) developed the consumer price index (CPI) a century ago to track the rate of inflation. BLS researchers fan out to stores and other businesses every month, recording the current prices of consumer goods. They use a mathematical formula to convert their data to a single number, based on a 1982–1984 benchmark of 100 points. A CPI above 100 indicates that goods cost more today than they did back then. A reading below 100 means that prices are lower now than they were in the early 1980s.
The beauty of the CPI is that we can use it to calculate the change in the cost of living over any period we choose. Consider this example: The CPI climbed from 168.8 points in January 2000 to 230.3 in January 2013.32 That’s an increase of 36.4 percent, which equals the inflation rate for that 13-year span.
The practical application is obvious. If you earned $50,000 at your job in 2000, your salary 13 years later would need to be 36.4 percent higher—$68,200—to give you exactly the same purchasing power. If you were paid more than $68,200 in 2013, you were ahead of the curve. If you were making less than that, you were just another victim of the rising cost of living.
Presidential politics has been hit with its own inflationary spiral, as we’ve already noted. If the CPI can effectively remove inflation as a factor when comparing prices or salaries, why can’t we use the same principle to equalize vote totals from different elections?
Well, that’s exactly what I’ve done. I have devised five sets of equalized election statistics, all easily identified by the suffix *EQ:
The same principle of equalization can be applied to all major aspects of presidential politics—primaries, conventions, and general elections. The formula, as you can see, is a simple one:
Equalized Votes = ((Candidate's Votes) / (All Votes)) • Base
Let’s go through the process step by step, using Jimmy Carter’s victory in the 1976 election as an example. We’ll focus on popular votes for the moment:
Step 1. Determine a candidate’s percentage of all votes cast. Round the answer to six decimal places. Carter drew 40,830,763 popular votes out of a total of 81,555,889. That yields a PV% of 50.064764 percent.
Step 2. Choose a base, which will be the size of your equalized electorate. Any figure could be used, but I have established 100 million as the base for all PV*EQ calculations. (More on that in a minute.)
Step 3. Multiply the candidate’s percentage and the base. Round the answer to the nearest whole number. Carter’s PV% (50.064764%) times the base (100,000,000) yields 50,064,764 equalized votes.
And that’s that. We can easily follow the same procedure to equalize Carter’s other statistics. His primary votes decline from an actual sum of 6,235,609 to an equalized 3,884,473. His convention total drops from 2,239 votes in real life to 744 after adjustment. And his electoral votes go from an actual 297 to an equalized figure of, well, 297.
What gives? Why do some of Carter’s vote totals increase after equalization, while others decrease or stay the same?
Look to the bases for the answer. The base for PV*EQ (100 million) is larger than the electorate in 1976 (81.6 million), so the equalization process inevitably expands the totals for that year’s candidates. But the base for FB*EQ (1,000) is smaller than the number of delegates to the 1976 Democratic convention (3,008), thereby triggering a reduction. And the size of the Electoral College is identical for the base period and 1976 (538), so there is no change at all.
Equalization will work with any base, no matter its size, but I’ve tried to come up with logical standards for the five categories. Here’s my thinking:
The base for a particular category never varies. An equalized total of 538 electoral votes, for example, has been allocated to every single election from 1789 through 2012. This uniformity allows us to make honest comparisons of presidential contenders throughout American history.
Equalization removes the chronological bias that elevated Barack Obama to first place on both popular-vote lists. Lyndon Johnson vaults to the top of the single-year rankings with 61.05 million equalized votes, while Obama’s 2008 performance sinks to 19th place. And Franklin Roosevelt, the only four-term president, soars to the top of the career standings, which is precisely where you would expect to find him. Roosevelt has a lifetime total of 226.28 million equalized votes. Obama is a distant 10th with 103.99 million.
We’ve made a decent start toward leveling the statistical field, yet that’s all it is—a start. Equalization allows us to compare the prowess of Woodrow Wilson and John Kennedy in Democratic primary elections, the success of Abraham Lincoln and Dwight Eisenhower at Republican conventions, and the strength of Andrew Jackson and Bill Clinton in the Electoral College. But it’s not an effective tool for every situation.
Let’s say we want to contrast the political records of a pair of two-term presidents, Thomas Jefferson and Ronald Reagan. It’s not easily done. We possess a full range of equalized stats for Reagan, but nothing more than EV*EQ for Jefferson, who reigned before primaries, conventions, and popular voting became the rage. Or maybe we’d like to assess the relative merits of a candidate who repeatedly failed to win the Republican nomination (such as Robert Taft) and a contemporary who consistently outperformed him (Tom Dewey comes to mind). Again not so simple. There’s a considerable difference in the quantity of data available for candidates who became nominees and those who fell short.
What we need is a single number that encapsulates the achievements of any presidential candidate—no matter the era, no matter the level of success. Remember WAR, the all-purpose baseball statistic? We need something like that.
I’ve developed three numbers that fit the bill in varying ways:
I’ll spare you the gory details, but I spent months developing these statistics. I ran endless comparisons of proposed indexes and the actual results of every election from 1789 through 2012. Eighteen PI formulas flowed through my computer, for example, before I settled on the one used in this book. Could further refinements be made? I’m absolutely certain they could. But I’m also pleased with the indexes as they stand. I believe they do what they’re supposed to do.
So let’s start with CS. Here’s the overall formula, followed by a breakdown of its five components:
CS = Qua + Nom + Gen + Min + Win
Qua: Anybody who reaches a qualification level earns 10 points. The variable x is 1 for any qualified candidate, 0 for anybody else:
Qua = x • 10
Nom: A major-party nominee is awarded 40 points, while each unsuccessful candidate for the nomination is given a lesser amount, based on his or her relative performance in the primaries and at the convention.
Nom1 is the relevant equation for nominees. The variable x is 1 for anybody who secures a nomination, 0 for the rest of the field:
Nom1 = x • 40
An equation with two components, Nom2, applies to any contender who falls short of a nomination. Component 1 involves the candidate’s total of equalized convention votes, either FB*EQ or LB*EQ, whichever is larger. That figure is divided by the number of equalized votes necessary for victory, 501. The Democrats required a supermajority of 667 equalized votes to secure a nomination until 1936, but the CS formula maintains a threshold of 501 for the sake of uniformity. If any unsuccessful candidate surpassed that total prior to 1936, 501 votes are entered in the equation.
Component 2 is based on PriV*EQ, which is divided by 5,000,001, the number that constitutes a majority of all equalized votes in the primaries. If any unsuccessful candidate drew a greater number of votes, 5,000,001 are entered in the formula.
The two components of the equation are multiplied, respectively, by y and z, which always add to 40. The values of these variables differ from era to era, as we’ll discuss in a bit:
Nom2 = (y • (FB*EQ or LB*EQ) / 501) + (z • PriV*EQ / 5,000,001)
Gen: Anybody who draws support in the general election is allocated points from a 50-point pool. The size of a candidate’s share is determined by the sum of his percentage of equalized electoral votes (which is multiplied by 30) and his percentage of equalized popular votes (which is multiplied by 20):
Gen = (30 • (EV*EQ / 538)) + (20 • (PV*EQ) / 100,000,000)
Min: Any independent or minor-party candidate who reaches a qualification level is eligible for as many as 30 points, based on equalized vote totals.
The variable a is 0 for any major-party nominee or any nonqualified candidate. It’s 1 for anybody else who receives votes in the general election. The equation takes the percentage of equalized electoral or popular votes, whichever is larger, and multiplies it by 300. If the relevant share of EV*EQ or PV*EQ is greater than 10 percent, the Min score is capped at 30 points. If the percentage is lower, the following formula is used:
Min = a • 300 • ((EV*EQ / 538) or (PV*EQ / 100,000,000))
Win: Any general-election winner is awarded a bonus that equals one-tenth of his or her Gen score.
Win = 10% • Gen
And there you have it—everything you need to know about the formula for campaign score.
Well, almost everything. There actually are six variants of the basic CS formula, reflecting slight changes between one period of American history and another. Adjustments are made whenever the stream of political data grows richer or moves in a new direction:
Franklin Roosevelt, Richard Nixon, and Ronald Reagan came close to sweeping the Electoral College in the 20th century, yet all three were limited to roughly 60 percent of the popular vote. It’s illogical to assume that Washington would have been any different. Yes, he was the transcendent hero of the Revolutionary War, but he was also a rich Southern Federalist. If common men had been allowed to vote in 1789, it’s likely that many of them—especially those who were poor, Northerners, or Anti-Federalists—would have preferred another candidate. And it’s hard to envision unreconstructed Tories being Washington supporters.
The omission of popular votes, fortunately enough, can be rectified by a simple equation. If we know a candidate’s EV*EQ, we can estimate how many popular votes he might have received:
PV*EQ = (50,543 • EV*EQ) + 32,705,460
I arrived at this equation by comparing PV% and EV% for all major-party nominees since 1824. It isn’t perfect, but it does generate a good ballpark figure for any candidate who ran in the first nine elections. Washington, for instance, is credited with 59.90 percent of the popular vote for each of his victories. That figure is entered into his CS formulas for 1789 and 1792.
Major-party nominees earn 40 Nom points in Versions 1 and 2, of course, but other candidates are shut out in the Nom column because of the lack of convention votes. This may seem unfair, though I don’t believe it is. All qualified candidates under Versions 1 and 2 received electoral votes, which wasn’t true of most aspirants after conventions arrived. The Gen scores for early candidates are consequently inflated, offsetting their shortfalls in Nom points.
The CS formula might be easier to understand if we run through a couple of examples. Let’s begin with the election of 1960, a tightly fought battle between John Kennedy and Richard Nixon:
Step 1. Kennedy and Nixon both reached qualification levels, so that’s 10 Qua points for each.
Step 2. Both men won major-party nominations, giving them 40 Nom points apiece.
Step 3. Kennedy defeated Nixon by a decent margin in equalized electoral votes (304 to 219), though his lead in equalized popular votes was razor-thin (49,720,535 to 49,548,285). Kennedy earned 16.95 points for EV*EQ (56.51% of 30 points) and 9.94 points for PV*EQ (49.72% of 20 points), for a total of 26.89 Gen points. Nixon received 22.12.
Step 4. Both men were major-party nominees, making them ineligible for Min points.
Step 5. Kennedy, as the general-election winner, received a Win bonus of 2.69 points, equal to 10 percent of his Gen score.
Here’s how their campaign scores add up:
CS = Qua + Nom + Gen + Min + Win
Kennedy = 10.00 + 40.00 + 26.89 + 0.00 + 2.69 = 79.58
Nixon = 10.00 + 40.00 + 22.12 + 0.00 + 0.00 = 72.12
Kennedy and Nixon ran conventional campaigns. They sought and secured major-party nominations, then faced off in November. But how does CS function for nontraditional candidates? Let’s look at the case of John Anderson, an Illinois congressman who initially ran for the Republican nomination in 1980, but eventually withdrew to become the presidential candidate of the hastily formed National Unity Party:
Step 1. Anderson drew a sufficient number of votes in the Republican primaries to become a qualified candidate and earn 10 Qua points.
Step 2. He did not win the Republican nomination, so Anderson’s Nom score was determined by the Nom2 equation. His 1,238,978 equalized primary votes worked out to 24.78 percent of the majority threshold of 5,000,001, thereby translating to 24.78 percent of 30 points, or 7.43 points. His 19 equalized convention votes accounted for 3.79 percent of the 501 votes needed for the nomination, yielding 3.79 percent of 10 points, or 0.38 points. His grand total: 7.81 Nom points.
Step 3. Anderson was shut out in the Electoral College, but his PV*EQ of 6,611,730 translated to 1.32 Gen points (6.61% of 20 points).
Step 4. He ran in the general election as a minor-party candidate, so Anderson was eligible to receive Min credits. His PV% (6.61%) was multiplied by 300 to come up with 19.84 points. (If you actually pulled out your calculator, you came up with 19.83 points, but 19.84 is correct. All percentages are rounded to two decimal places in this book, but are carried to additional digits when calculations are made.)
Step 5. Anderson finished third in the general election, so he did not earn the Win bonus.
Here are the results:
CS = Qua + Nom + Gen + Min + Win
Anderson = 10.00 + 7.81 + 1.32 + 19.84 + 0.00 = 38.97
CS is a remarkably good indicator of a candidate’s performance in a given year. General-election winners typically land between 75 and 100 points, ranging from landslide victors in the 90s to nail-biters in the upper 70s. Major-party nominees who lost general elections are grouped from the 50s to the middle 70s. They’re followed by significant minor-party candidates and strong (but ultimately unsuccessful) contenders for major-party nominations, who can be found in the 30s and 40s. The remaining qualified candidates—those who had little, if any, impact—are lumped between 10 and 29 on the scale.
CS is the single number we’ve been looking for. It enables us to make comparisons that previously were impossible, such as the matchups I suggested at the beginning of this section. We can now say that Ronald Reagan’s best year (95.13 points in 1984) surpassed Thomas Jefferson’s top performance (93.05 in 1804), but Jefferson’s cumulative CS score (271.57 in four elections) was slightly higher than Reagan’s career total (262.11). We’re also able to estimate that Thomas Dewey was twice as successful as his rival who also sought the Republican nomination three times, Robert Taft. That’s because Dewey’s career CS of 175.94 was almost precisely twice as large as Taft’s 87.14.
There’s more to be said about CS, but we’ll save it for later. It’s time to move on to the other two indicators that I mentioned earlier—PI and ROP.
The dictionary defines potential as “a quality that something has that can be developed to make it better” or “an ability that someone has that can be developed to help that person become successful.” The key phrase in both of those definitions—“can be developed”—makes it clear that potential, by itself, is no guarantee of victory. If a candidate doesn’t aspire to success, doesn’t work hard enough, and doesn’t benefit from an occasional lucky break, his or her potential is unlikely to be fulfilled.
American history is replete with politicians who appeared formidable on paper, yet were unable to rendezvous with destiny. They failed to develop their qualities and abilities.
Adlai Stevenson, the former governor of Illinois, remained an attractive political personality despite losses to Dwight Eisenhower in the elections of 1952 and 1956. Every Gallup Poll from the start of 1959 found Stevenson running neck and neck with John Kennedy for the following year’s Democratic nomination.33 Pundits noted that William Jennings Bryan had been tapped a third time by the Democrats despite two general-election defeats. Why not Stevenson?
Why not, indeed? My calculations show that Stevenson entered the 1960 campaign with a potential index of 7.2 on a 10-point scale, putting him ahead of every Democratic contender except Lyndon Johnson. He stood a full point in front of Kennedy. If the battle for the nomination had eventually narrowed to a two-man duel, their respective PI scores gave Stevenson a 61 percent chance of defeating Kennedy.
But there was an essential difference between the two candidates. Stevenson was famously indecisive. A story made the rounds in 1960 about one of his speaking engagements. It was a fictional tale, yet it carried the ring of truth. Stevenson, so the story went, was told he had just five minutes before his turn at the podium. “Do I have time to go to the bathroom?” he asked an aide, who assured him he did. That merely led to another question: “Do I want to go to the bathroom?”34
Kennedy, on the other hand, was a driven man. He was still three months shy of his 43rd birthday when he hit the campaign trail in the key primary state of Wisconsin. An elderly woman approached him, making no effort to hide her skepticism. “You’re too soon, my boy, too soon,” she said. Kennedy smiled back at the old lady. He had been told the same thing by people of much greater stature, and he would not be dissuaded. “No, this is my time,” he replied. “My time is now.”35
Kennedy, of course, kept pushing until he converted his potential into reality, winning the Democratic nomination and then the presidency. Stevenson hemmed and hawed for much of the year, eventually mounting a halfhearted and predictably unsuccessful campaign. He wound up working for his younger rival as ambassador to the United Nations.
It’s impossible to measure spirit, intensity, and other intangible qualities, but several aspects of a prospective presidential candidacy can be quantified. History tells us that the odds of reaching the White House are better for a senator or governor than a congressman or cabinet member, better for a candidate in his 50s than somebody in his 30s or 70s, and better for a resident of California or New York than a person who hails from Utah or Wyoming.
My potential-index formula encompasses nine quantifiable factors. I based the scores for each component on my analysis of election returns from four distinct eras, cleverly identified here as Era 1 (1789–1836), Era 2 (1840–1908), Era 3 (1912–1956), and Era 4 (1960–present). Trends varied from era to era, necessitating adjustments in my point scale. Cabinet members, for example, were extremely successful as presidential candidates in Era 1, but steadily declined in influence thereafter. Midwesterners outperformed contenders from other regions in Era 2, only to be supplanted by Easterners in Era 3 and Southerners in Era 4. And the optimal age for a presidential candidate oscillated from 60 in the initial era to 54, 57, and 53 in the periods that followed.
The potential index is a flexible indicator, applying not only to the past but also to the future. Chapters 2–5 contain PI ratings for all candidates from 1789 to 2012. The same system can be used to analyze 2016’s contenders. (Go to www.countingthevotes.com to see their PI scores.) Here is the overall formula, followed by an explanation of its components:
PI = Siz + Reg + Age + Job + Maj + Sta + Fed – Min – Rep
Siz: A home state is the place where an individual maintains a voting address during his presidential candidacy, not necessarily the state where he was born. (A brief interjection: I applaud gender equity, but I won’t be using “he or she” every time I lapse into the third person. There have been 289 qualified candidates in the history of presidential politics, and all but three were men, so the masculine pronoun naturally predominates. No offense intended.) A candidate’s score is based on the size of his home state, as reflected by its representation in the Electoral College. Use the line in the following chart that begins with the appropriate whole number. If a home state wields 1.8 percent of all U.S. electoral votes, for example, the whole number is 1. Follow the 1% line across to the era in question. The maximum Siz score is 1.50 points:
Reg: Up to 1.50 Reg points can be awarded to a candidate, depending on the region in which his home state is located. Regional identities are obvious in most cases, so I won’t bother with a complete list. I can think of only eight states where there might be some confusion. Delaware, Maryland, West Virginia, and the District of Columbia (which is treated as a state in presidential elections) are in the East; Kentucky, Oklahoma, and Texas are in the South; Missouri is in the Midwest:
| Region | Era 1 | Era 2 | Era 3 | Era 4 |
| East | 0.69 | 1.13 | 1.50 | 0.72 |
| South | 1.50 | 0.32 | 0.17 | 1.50 |
| Midwest | 0.05 | 1.50 | 0.74 | 0.81 |
| West | 0.00 | 0.06 | 0.24 | 0.80 |
Age: Election Day as we know it—the Tuesday following the first Monday in November—wasn’t mandated until 1845.36 Previous elections were staggered over several days or even weeks. That posed a problem, since I wanted to calculate the ages of all candidates the same way. I finally chose a standard benchmark of November 1 in the appropriate year, no matter what month the election actually took place. The highest possible Age score is 1.50 points:
Job: This is the most important factor in the PI formula, accounting for a maximum of 3.00 points. Job scores are assigned according to a contender’s current or most recent title, selected from the following chart of 21 positions. If a candidate held another job of greater value during the eight years preceding an election, he or she receives the higher score. (This stipulation will affect the 2016 potential indexes for two prominent senators who became cabinet members, Hillary Clinton and John Kerry.) An element of subjectivity is occasionally needed to fit a candidate into the list of 21 positions. Anybody involved in business, for instance, is classified as an industrialist, and any political activist who has never held public office is listed as a movement leader:
| Position | Era 1 | Era 2 | Era 3 | Era 4 |
| Ambassador | 0.36 | 0.03 | 0.18 | 0.09 |
| Cabinet member | 1.95 | 1.11 | 0.51 | 0.00 |
| College president | 0.00 | 0.00 | 0.06 | 0.00 |
| District attorney | 0.00 | 0.00 | 0.03 | 0.00 |
| Federal judge | 0.00 | 0.03 | 0.00 | 0.00 |
| General | 0.57 | 1.53 | 0.45 | 0.03 |
| Governor | 0.42 | 2.70 | 1.98 | 2.70 |
| Industrialist | 0.00 | 0.03 | 0.21 | 0.06 |
| Journalist | 0.00 | 0.24 | 0.06 | 0.12 |
| Mayor | 0.21 | 0.00 | 0.03 | 0.06 |
| Movement leader | 0.00 | 0.09 | 0.09 | 0.21 |
| President | 3.00 | 2.85 | 3.00 | 3.00 |
| Presidential nominee | 0.21 | 0.66 | 0.36 | 0.57 |
| Representative | 0.03 | 1.44 | 0.18 | 0.45 |
| Senator | 1.77 | 3.00 | 1.35 | 2.79 |
| State judge | 0.03 | 0.15 | 0.03 | 0.00 |
| State official (not governor) | 0.09 | 0.03 | 0.00 | 0.00 |
| Supreme Court justice | 0.09 | 0.12 | 0.15 | 0.00 |
| Territorial delegate | 0.00 | 0.03 | 0.00 | 0.00 |
| Vice president | 1.62 | 0.24 | 0.12 | 1.05 |
| Vice presidential nominee | 0.00 | 0.00 | 0.00 | 0.06 |
Maj: A candidate who ran on a major-party ticket in a previous election, either as a presidential or vice presidential nominee, gets a Maj score of 1.50. Everybody else receives 0.60 points.
Sta: Anybody who has ever been elected to either of the two major statewide offices—governor or U.S. senator—is awarded 0.50 Sta points. Those candidates who lack such experience are given 0.22 points.
Fed: A previous job in the federal government can earn a candidate a Fed score of 0.50. Several positions meet this requirement, including president, vice president, senator, cabinet member, ambassador, federal judge, general, or key presidential aide. Anybody without a federal background receives 0.09 points.
Min: This is the first of two deductions. Most candidates who represent minor parties have 1.50 Min points subtracted from their totals. An exception is made for anybody who begins the election year seeking a major-party nomination before shifting in midstream to a minor-party ticket, such as the previously mentioned John Anderson. These party switchers do not lose Min points.
Rep: Repeat candidates can also be subject to a deduction. If a contender was a qualified candidate in two previous elections, he loses 0.75 points on his third try. If his prior number of qualified candidacies climbs to three or more, he loses a total of 1.50 points. Incumbent and former presidents are exempt from this rule.
Scores for the nine factors are combined, then rounded to a single decimal place. Let’s quickly run through the steps, using Thomas Dewey’s 1948 campaign as an example.
Step 1. Dewey’s home state, New York, cast 47 electoral votes in 1948, accounting for 8.85 percent of the national total of 531. Go to the Siz table, drop down to 8% in the left-hand column, and scan across to Era 3 (1912–1956). The result: 1.50 points.
Step 2. New York is in the East, which earned Dewey a Reg score of 1.50.
Step 3. Dewey was 46.61 years old as of November 1, 1948. Era 3 candidates of that age received 0.30 points.
Step 4. The charts in this book show Dewey’s occupation in 1948 as “former nominee.” That’s because most Americans knew him best as the Republican standard-bearer in the previous election. But he was also the governor of New York (1.98 Job points), which trumped his status as a former nominee (0.36 Job points).
Step 5. Dewey received 1.50 Maj points for having been on the 1944 Republican ticket.
Step 6. His experience as a statewide official brought Dewey 0.50 Sta points.
Step 7. Dewey’s resumé did not include any of the positions that earn full credit in the Fed category, so he received the consolation prize of 0.09 points.
Step 8. There was no Min deduction, since Dewey ran as a Republican.
Step 9. Dewey’s 1948 campaign was his third qualified candidacy, coming on the heels of unsuccessful efforts in 1940 and 1944. A third candidacy carries a deduction of 0.75 Rep points.
And the results:
PI = Siz + Reg + Age + Job + Maj + Sta + Fed – Min – Rep
Dewey = 1.50 + 1.50 + 0.30 + 1.98 + 1.50 + 0.50 + 0.09 – 0.00
– 0.75 = 6.6
There is a strong correlation between a candidate’s potential index and his chance of success. The trend becomes obvious when we break down the 444 qualified candidacies between 1789 and 2012, dividing them into four PI ranges. (Don’t forget this small, yet important, distinction: There have been 289 qualified candidates in American history, but roughly one-third of them ran more than once. Hence 444 qualified candidacies.) The count for each group is followed by the percentage of contenders who became major-party nominees and general-election winners:
| Group | PI range | Qualified | N% | W% |
| Group 1 | 8.0 or more | 60 | 56.7% | 41.7% |
| Group 2 | 6.0-7.9 | 182 | 26.9% | 13.2% |
| Group 3 | 4.0-5.9 | 143 | 19.6% | 5.6% |
| Group 4 | 3.9 or less | 59 | 3.4% | 0.0% |
We can use a nominee’s potential index to estimate his chance of victory (CV%) in any general election, as I mentioned a few pages back. It’s a simple two-step procedure. I have derived a pair of formulas that define the link between PI and success more precisely than the previous chart, based on an analysis of all 57 presidential elections.
We begin by determining the raw CV score for each major-party nominee. (If a minor-party candidate has a realistic chance of winning, he should also be included.) These numbers mean nothing in themselves, but they do indicate the relative strength of the contenders. Any candidate with a potential index of less than 5.0 is assigned a raw CV of 0.0100. This formula is used for anybody with a PI of 5.0 or greater:
Raw CV = (0.0163 • PI2) - (0.0995 • PI) + 0.1023
Each candidate’s raw CV score is then divided by the sum for all nominees, yielding the estimated chance of victory. Here’s the calculation for a typical head-to-head matchup:
CV% = Raw CV1 / (Raw CV1 + Raw CV2)
Let’s slip back to 1948 to put these formulas into action. Only two candidates had any hope of winning the general election that year—Harry Truman (with a PI of 7.6) and Thomas Dewey (6.6). These are their raw CV scores:
Truman Raw CV = (0.0163 • 7.62) — (0.0995 • 7.6) + 0.1023 = 0.2876
Dewey Raw CV = (0.0163 • 6.62) — (0.0995 • 6.6) + 0.1023 = 0.1556
We then plug these numbers into the second formula to estimate the chance of victory for each nominee:
Truman CV% = 0.2876 / (0.2876 + 0.1556) = 64.9%
Dewey CV% = 0.1556 / (0.2876 + 0.1556) = 35.1%
If a nominee with a potential index of 7.6 confronts an opponent with a PI of 6.6, history suggests that the former will win about 65 percent of the time. Truman’s victory, as I previously noted, should not have come as such a surprise.
One other indicator, return on potential, also mentioned earlier, deserves a few words here. ROP determines whether a candidate fulfilled his promise in a given election. The underlying principle is the same one followed by portfolio managers when they calculate whether they’re receiving a suitable return on their investments.
The ROP formula takes a ratio of performance (CS) to potential (PI), multiplies it by 10, and rounds it to the nearest whole number:
ROP = 10 • (CS / PI)
If a candidate’s ROP exceeds 100, he or she has done better than could have been expected. A two-digit score, on the other hand, is a sign of underperformance.
A brief return to 1960 dramatizes the difference. You’ll recall that Adlai Stevenson held a PI advantage over John Kennedy, 7.2 to 6.2. But Kennedy won the election with a CS of 79.58, with Stevenson lagging far behind at 12.86. Kennedy, who never stopped pushing for success, posted an impressive ROP of 128 points. Stevenson, who never truly decided whether he wanted to be president, finished with an abysmal ROP of 18.
There, we’ve covered the basics.
This chapter has unveiled several new concepts—qualified candidates and equalization, for instance—as well as an array of definitions and abbreviations. I have delineated 12 basic statistics from PriV to EV, along with five equalized counterparts from PriV*EQ to EV*EQ. I have described the inner workings of three specialized indexes: CS, PI, and ROP.
These are notions and numbers that you’ll encounter throughout the book, so the only logical step was to clarify them at the start.
But a wide range of ideas and statistics have not yet been introduced— scenarios, flips, simulations, average state deviations, PI gaps, EV-PV gaps, and bellwether streaks, among others. Most are confined to a single part of this almanac, so I’ll explain each in its appropriate chapter.
These indicators focus on different aspects of the electoral process, yet they are tied together by a common numerical thread, the strong belief that statistics can help us to unravel the deep mysteries of presidential politics. “When you can measure what you are speaking about, and express it in numbers, you know something about it,” said William Thomson, the great physicist and mathematician who was ennobled as Lord Kelvin. “But when you cannot measure, when you cannot express it in numbers, your knowledge is of a meager and unsatisfactory kind.”37
It’s time for us to do our own measuring. Let’s start at the very beginning.