As the melodrama unfolded, each candidate's ups and downs were tracked not just in the polls but, for the first time in modern history, by the prediction markets.

Prediction markets are, for lack of a better term, just betting lines. Bettors can bet on who they think will win, by how much, and in which states. On Polymarket, which only accepts cryptocurrency bets, bettors could even have bet on whether Taylor Swift would attend Vice President Kamala Harris' rally on Monday. (She went to the Kansas City Chiefs game instead.)

The central question of whether Harris or former President Donald Trump would win the presidential election preoccupied Polymarket with $3.2 billion in bets on Election Day. Most bets were on Trump, with $1.3 billion placed on the former president and $827 million placed on Harris. Polymarket did not respond to a request for comment.

The rise of betting markets signaled a significant shift in the way elections were reported in the press and observed by the public. Pundits, analysts, and Wall Street investors began incorporating prediction markets into their election predictions. The reason for this was that they reflected people's true feelings because they were putting their own money at risk. In reality, this was only partially true: all betting lines fluctuate due to both betting volume and the external factors affecting the race.

In late July, for example, Trump's chances skyrocketed following President Joe Biden's abysmal performance in the debate. However, according to Bloomberg, the spike in Trump probabilities in late October was due to divergent betting behavior by just 1% of Polymarket users.

The financial world was as attached to the prediction markets as politicians were. “Various assets were impacted by Polymarket’s results,” said Lukman Otunuga, senior market analyst at online trading platform FTXM Assets. “This was reflected in global markets as the 'Trump trade' boosted the US dollar, Bitcoin and financial stocks in October. It appears investors have used the prediction markets to gain an advantage ahead of Election Day.”

There appear to be signs in the final days of the campaign that some Trump contests are weakening, in part because the volatility of the former president's electoral prospects suggested he could lose after being favored for much of the race.

In this election cycle, the volatility of the prediction markets counterbalanced the stability of the polls. Traditional voter polls continued to show a statistical dead heat, while betting markets periodically fluctuated — swinging because of Harris' strong debate performance, Trump's recent offensive comments and, yes, even late, dramatic polls from star Iowa pollster Ann Selzer.

For some bettors (or are they investors?) the volatility is the appeal. In prediction markets, bettors purchase contracts that pay out $1 if the outcome they bet on occurs. The price of this contract changes with the probability of the outcome. But bettors can also sell contracts on these platforms. So if you time the market correctly, you can theoretically profit even if the outcome you bet on doesn't happen. For example, on July 21, the day Biden dropped out, a Harris victory cost 29 cents. A week later, the price was 39 cents – a spread large enough to make a significant and guaranteed profit if one had bought low and sold high.

But for all the truth-telling abilities of the market claimed by prediction market proponents, they can still stumble: People make bad investments all the time. A heavily favored Trump could lose, or the delicious gains of an underdog bet on Harris may never materialize.

Meet business's brightest minds and boldest leaders at the Fortune Global Forum, November 11th and 12th in New York City. Inspiring sessions and intimate discussions include Fortune 500 CEOs, former cabinet members and global ambassadors, as well as seven-time world champion Tom Brady.

View the full agenda here or request your invitation.