US Election: How previous presidential elections panned out during the final 100 days

Exactly 100 days remain until the US election - a useful marker for the beginning of the final stretch. While we await official confirmation at the party conventions, it would require something extremely dramatic to prevent Donald Trump and Joe Biden being the nominees.

Current signals point strongly towards Biden, whose odds are steadily shortening on the exchange. The former Vice President is a 1.618/13 chance compared to 2.8615/8 about Trump.

There is, however, plenty of time for change. Whatever the polls and odds say, plenty of punters will be prepared to dismiss those indicators to back the underdog after Trump's stunning turnaround last time.

How does Biden's position measure up against past presidential races? This will be the fifth US election ever to be traded on Betfair. Here's what happened previously.

2016: Shortest ever 100 day favourite goes on to lose

In many respects, trading at the 100-day stage was typical of that extraordinary 2016 race - a landmark in the history of political betting.

The odds fluctuated considerably throughout the day - the average odds matched were 1.43 for Clinton, 3.25 for Trump. More than ten times as many bets were placed as the equivalent day in 2012. Despite being clear outsider, two thirds of those bets were placed on Trump. Further evidence of his gamechanging effect on politics.

History will of course record that the market was wrong. Clinton was the strongest of any favourite at this stage during the Betfair era - considerably more so than Biden despite a smaller poll lead. In the last seven polls completed in July, listed on RealClearPolitics - Clinton led by an average below 2%. Biden's current RCP average lead is 8.7%.

Therein lies a great myth of 2016. The polls weren't so wrong. Rather it was we pundits and punters that over-rated Clinton. She would maintain a solid overall lead on both indicators but Trump did lead occasionally and there were blips.

When Clinton collapsed with pneumonia at a 9/11 memorial, her odds would drift towards even money without ever ceding favouritism. Going into the first debate, she was around 1.68/13. By the second, following Trump's humiliation by the 'Pussygate' tapes, she was into 1.282/7. The broader trend soon returned though, with the last 11 polls showing only one Clinton lead above 4% (she won the popular vote by 2.1%).

2012: Obama proves the market right

In terms of drama and unpredictability, 2012 was the direct opposite of 2016. Indeed it goes a long way towards explaining why there was such confidence in Clinton's poll lead four years later.

Barack Obama was 1.635/8 to win a second term with 100 days to go, compared to 2.7 about Mitt Romney. He would never cede favouritism en route to a 3.9% victory - in line with his late July average but notably better than the 0.7% final RCP average. It was argued that the polls couldn't weigh the superior Democrat ground game and turnout operation.

The only blip came when Obama was widely acknowledged to lose the first debate. Romney briefly took a small lead in some polls but that may have actually helped the incumbent. Obama upped his game in the two subsequent debates and any hint of complacency soon disappeared amid a media narrative that it was 'too close to call'.

In fact, state polls consistently pointed towards an electoral college win and these drove market trends more than nationwide figures. If memory serves, collating and applying those numbers proved the making of Nate Silver's reputation. Obama steadily shortened in the betting as polling day neared, trading around 1.2 on election morning.

2008: Wider conditions make Obama's election very predictable

Trump fans won't want to hear it but, right now, 2008 has the closest parallels with 2020. Republican incumbent George W Bush left office with very low approval ratings, against a dire economic backdrop. The financial crash and subsequent bailout of banks and insurance companies would leave a near-impossible task for GOP successor John McCain.

One hundred days out, Barack Obama was trading at 1.4640/85 compared to 3.55/2 about McCain. His lead was around 4%, which he would extend to 7.3% in November (almost bang in line with polls).

As in 2012 and during the Trump presidency, the blue/red partisan split seemed largely immovable. McCain never looked like winning. He would gamble on Sarah Palin as running mate to enthuse the base. Whilst achieving that, her largely derided candidacy would further alienate the critical independent voters.

2004: Bush victorious after exit poll skews the betting

The 2004 election was a landmark as the first US election on Betfair. It was also mine and a real eye-opener into how exciting these markets can be. At the 100-day stage the betting was perfectly even - both George W Bush and John Kerry traded at odds-on that day. It would swing back and forth, whilst remaining an unpredictable contest right up until the final hours.

At this stage, Kerry held a small average lead around 1%. By the end of August, Bush was ahead and stayed there. His victory margin was 2.4%, compared to a 1.5% RCP average.

It may therefore surprise readers to learn that, on the night, odds up to 5.04/1 were available about Bush. A massive over-reaction based seemingly on one exit poll predicting Kerry.

The mismatch helped fuel a narrative that betting markets were biased to the Left, which some argue was the case in 2016. That Democrats struggle to get their vote out while Republicans are less willing to complete polls or admit their allegiance. Biden-sceptics believe that will be the case in 2020. Time will tell whether he proves to be another Kerry, Obama or Clinton.

Follow Paul on Twitter and check out his website, Political Gambler.