What a Difference a Day Makes
In the annals of U.S. history, there are few moments which live up to Election Day, November 8 2016 for the sheer number of experts it left dumbfounded.
And in the new internet tradition of “photos taken right before something bad happened”, the photos of Hillary Clinton and her supporters smiling and exultant, mingling at her ill-fated victory party on election night, and at campaign events in the days leading up to her defeat, rank fairly high as well.
In 2016, everyone who was anyone thought Hillary Clinton was a shoo-in. Not least of which were Republicans and Donald Trump himself.
All the polls predicted Hillary Clinton would win by a landslide; the betting odds heavily favored Clinton. Everyone in the media was convinced Hillary Clinton was going to be the first female President of the United States.
When did it become clear to members of the press that Hillary Clinton was losing the election to Donald Trump?
Between 8:00–8:30 p.m. on election night, the truth started sinking in for some. For other journalists it was after 10:00 p.m. that night, long after any hope of Clinton pulling-off an electoral victory were long since dashed by the votes pouring in for Donald Trump.
The level of certainty in Trump’s defeat wasn’t confined to the U.S. Far from it; people all over the world like to bet as much as anyone else. Hillary Clinton was the odds-on favorite everywhere.
One British Trump fan won $124,000 in 2016 betting against the betting house oddsmakers. Using his own system of analytics and data observation, John Mappin predicted Donald Trump’s series of surprising wins only three days after Trump announced his candidacy.
When John Mappin bet on Trump to win the Republican primary, he’d never placed a bet before in his life. Mappin bet again on Trump to win the general.
Betfair, the betting house with which Mappin placed his long-shot bets, put Hillary Clinton’s odds of beating Donald Trump at 80%. After Donald Trump’s upset, the company lost millions.
John Mappin wasn’t the only gambler willing to take a long-shot on a political newcomer like Donald Trump. Betfair paid one Trump bettor $2.5 million.
Other U.K. betting companies put Clinton’s odds of winning even higher; Ladbrokes estimated a Clinton victory was 85% likely. Giving Donald Trump an only 15% chance of winning meant bets on him paid off six-fold. Ladbrokes lost big money on the U.S. election in 2016, too.
There was a moment in 2016 when oddsmakers caught on and Trump became the heavy favorite to win the presidency: On Election Day, Tuesday November 8 2016, at 10:57 p.m.
At that point, according to the betting market aggregator, Donald J. Trump became the 95.3% favorite to win the presidency. This, of course, was an enormous reversal in his predicted fortunes; leaping 77 percentage points in a single bound.
The New York Times had to revise its prediction in a similar fashion, and as suddenly. The paper, which had favored Clinton by between 85% and 91%, changed completely as the electoral math shifted tectonically to make Trump the favorite by 94%.
Of course, once the votes come in, anyone can predict with reasonable certainty who will be elected. Hindsight is 20/20; professional gamblers who lost money betting on Clinton in 2016 could have made a fortune betting on Trump, against the oddsmakers who got it so wrong.
Also on the other side of the pond, but on the other side of aisle, a record-breaking bet was placed on Hillary Clinton to win the presidency. The $615,862 bet would have netted a profit of around $224,000.
“As far as we are aware, this is the biggest wager ever taken on the US Presidential Elections. As a result of taking this bet, a win for Donald Trump has now become the favored outcome for us in terms of our liabilities for either of the two leading contenders.” — William Hill spokesman, William Hill. Time- Oct. 7, 2016
The betting firm William Hill might have gotten lucky; other firms took a beating.
This time around, you can bet betting firms and oddsmakers are going to make sure that whatever mistakes caused them to overestimate Clinton’s chances in 2016 don’t blind them in 2020.
Bias against Donald Trump is perfectly acceptable if you are a journalist reporting on his reelection prospects. Bias against Donald Trump might be perfectly acceptable in a pollster hired by the Democratic Party, or CNN, to get a certain poll result. For pollsters and the press, bias against Donald Trump is a factor; and it's a big one.
Betting firms have no use for bias against Donald Trump and it plays no role whatsoever in their calculations.
Oddsmakers can’t afford to let personal political preferences or their feelings about Donald Trump interfere with a sober analytical treatment of the data and an accurate prediction of the outcome based on that data.
Looking back at 2016 is useful in noticing parallels for 2020.
Pollsters and the press seem just as certain that Joe Biden will beat Donald Trump as they were about Clinton in 2016. They seem less certain about what went wrong with polling in 2016 and how it has been improved since then.
Betting oddsmakers, on the other hand, are in the business of making money. That they are putting that money on Donald Trump in 2020, after getting it so wrong in 2016, should worry Democrats, pollsters and the press far more than it does.
(contributing writer, Brooke Bell)