Election 2020: A look at the odds - a bookie’s nightmare

Trump vs Biden: A bookie’s nightmare

Bookies have been offering outrights on the 2020 US election for years, with the market opening immediately after President Donald Trump’s against-the-odds victory over Hillary Clinton four years ago.

On the eve of Election Day, betting pundits are rightly asking if Trump can do it again.

There are some obvious similarities to four years ago: While we know much more about him than we did four years ago, Donald Trump is still Donald Trump (though if we’ve learned anything about him in his first term in office, it’s that he can still be as unpredictable as his late night Tweets). And his opponent is, once again, a high-profile established Washington lifer.

The other thing that hasn’t changed are the nightmares that Trump is still giving bookmakers - which is precisely the subject of this article.

To set the stage, some background. Outright markets in the early stages had a diverse array of candidates, and it was only after Trump and Biden secured their respective parties’ nominations that the field was whittled down to our main protagonists - effectively shutting the door on bettors’ hopes for long-shot candidates (though you can still get a bet on Jo Jorgensen).

What shapes the odds?

Initial odds are formed on presidential candidates by bookies in pretty much the same way as a major sporting event: The book takes a view on the chances of each candidate winning, splitting up the 100% quantum of probabilities between the number of candidates left in the race.

In politics betting, that view might be influenced by things like how healthy or mentally sharp a candidate is, or looks; how much relevant experience they can bring to the role; what skeletons are in (or in some cases, out of) their closet; how influential they are in galvanising their parties around them, or how influential they could be in galvanising communities within the country. The book distills these factors into its own view on anticipated investment and liability controls - i.e., how much exposure is it willing to take on any one candidate.

In short - as with a sporting contest - what are the chances of A prevailing over B, what influences those chances, and what price would you charge someone to back those chances. Really, it boils down to what’s the price at which a book would be willing to accept the risk of losing a bet?

Not enough data, captain

But the major obstacle when it comes to setting prices for elections is a lack of observable data or analysis. Yes, there are a plethora of polls and more news and rumour than you can swing a cat at - but at the end of the day it all boils down to one contest - one contest that has never been played before.

Yes, Trump has won one election; and in Biden’s case, two - though they were both as Barack Obama’s vice president. He’s never competed for the top job. This time, it’s Trump versus Biden - for the first, and likely only, time. Which means bookies don’t know, really, what variables matter in this instance because there is no historical experience or data to go by.

To illustrate, contrast this with the number of times Manchester United might face Arsenal in a season - or, to equate this to a US election cycle, across four years. Caveating out the Covid uncertainties, oddsmakers have millions of data points to use, but more importantly, they have a rock-solid understanding of which of those data points really matter most for a particular contest: because they’ve seen it before, numerous times.

Let’s accept then that prices on the US election will have to be taken with a great dollop of good faith, and are more a vote of trust in the bookmaker setting the odds.

Next, is how bettors respond to those prices based on what influences their perceptions of value and candidate probabilities. And the quantum of bets they place ostensibly is a good safe indicator for bookies to shape the prices for either candidate. We say “ostensibly” with good reason...

In polls, we trust

In any election, the polls certainly can sway voting intentions - and betting odds. Current polls can be viewed as a reflection of all that is currently known, or perceived, about a candidate’s ability to win an election. In the absence of any substantive knowledge or ability to predict the future, a poll is the best guide, the “best real-world estimate” of either candidate’s chances of victory.

If you can trust the polls, you can simply equate a reliable poll percentage with an implied probability, and you have a reliable price indicator - right?

Wrong - as anyone who followed betting odds during the 2016 election will tell you. Prices on Trump offered by some books in the lead up to Election Day had widened to as far as 9 (+800) - implying a win probability of 11%.

And we know what happened - a shock result, and surely some very happy punters to boot.

Indeed, it is precisely the reason that polls in the past have proved so inaccurate that the Cloudbet Blog’s Election 2020 coverage has focussed on the idea that - or at least, asking the question if - betting markets may hold more accurate predictive power than the polls.

So here then is the second major challenge for bookmakers: Trump has beaten the odds before. Do you really want to expose your business to massive losses by pricing him too far out of the market, whatever the polls are telling you?

Where’s the action?

Bettors are clearly aware of what happened in 2016 - and books are clearly seeing substantial interest on Trump, in spite of what the election polls indicate. More than 85% of the money that Cloudbet has taken on the US election has been on Trump and the Republicans - and it’s a theme evident at other books as well.

While there could be a point made about the predispositions within the bitcoin community that might make them more right-leaning (if not directly Trump-supporting), people are certainly betting that he could do it again.

That being said, why is Biden still favoured to win by every book on the street? Someone must be betting on him, and in size. While we can’t speak to their motivations, the point we can make from a pricing standpoint is that it pays to be aware of where the action is, and who has it.

We estimate that betting exchanges are seeing roughly half of the action on the US election, with Betfair clearly taking some sizable Biden bets. We’re not suggesting that exchanges are the key indicator of an event’s outcome - but from a bettor’s perspective, it’s good to know where the action is to form individual views on what value you see in the odds.

The outright winner market

At the time of writing, the market at Cloudbet offers implied probabilities of winning at 66.6% for Joe Biden and 35.4% for the president.

Who will win in the end?

We can’t say for sure, but we hope that with this article, you’re better equipped to understand what goes into these odds.

What we can say is this: where Donald J. Trump is concerned, anything is possible.