As recently as 2014, political betting was arguably the most reliable market for favourite-backers. Suddenly it's become a haven for historic upsets. Anyone who thought 2016 was a freak year got a rude awakening last month. Based on their respective positions at the start of each campaign, Labour denying a Conservative majority was a bigger upset than either Brexit or Donald Trump's defeat of Hillary Clinton.
Market after market was turned on it's head. The reason, of course, is that expert opinion, narrative and therefore markets are driven by conventional wisdom - which breaks down during changing times. So what is changing and what lessons can we learn going forward?
The 'late swing towards the status quo' theory is bust
It used to be a given that governments would recover late in the campaign, as voters stuck with 'the devil they know'. It sounds logical and the evidence stacked up. The theory applied to every Conservative win between 1983 and 1992, the 2014 Scottish Independence referendum and when the last three incumbent US presidents won a second term. It was probably a factor behind David Cameron's surprise majority in 2015.
After a year of earthquake election results, the theory no longer inspires confidence. We will never know whether there was a late swing to the status quo in the EU referendum - as implied by the polls - because it is quite possible that the polls were always understating Leave due to differential turnout.
However the theory certainly fell down with Trump, as undecided voters broke decidedly for the outsider candidate and that phenomenon very much seems to apply to Jeremy Corbyn. While it was no great surprise to see a pro-Labour swing mid-campaign, their late momentum was unprecedented in living memory. Only one poll during the entire campaign predicted they would get 41%. Indeed, their 40-45% vote share was matched well above 200 to 1 on our exchange!
Campaigns can transform entrenched perceptions of leaders
It has long been argued that poll numbers for 'Best PM' and 'economic competence' were more reliable indicators than headline voting intention. The bare result supports the theory but again, the wider momentum points to a rapidly changing political environment.
Historically, such image and competence ratings tended to become entrenched long before the campaign. That plainly isn't the recent case. The transformation between May and Corbyn's ratings was unprecedented during a campaign and has gathered pace since the election. The last two months have shown that voters have never been more volatile, and how swiftly party fortunes can change under a different leader. Remember that before assuming the Tories current travails are irreversible.
Younger and previous non-voters can no longer be dismissed
All recent earthquake results can be explained in substantial part by differential turnout. Polling models overstated Labour turnout in 2015 and badly missed the difference in motivation between Remain and Leave voters. Trump's electoral college victory was fundamentally due to a combination of low turnout among likely Democrat voters in a few key states and Trump reaching new voters.
In each case, the demographic divide between older conservatives and younger liberals proved decisive. This election bucked the trend, blindsiding most pollsters, the commentariat and betting markets. Yougov and Survation's widely mocked high turnout models fared best.
So is this the new normal? Could it be that those historic defeats have awakened people to the importance of voting - seeing how it could literally change the world? Perhaps, as I suggested during the election when advising a bet on higher turnout, Brexit has enlivened debate and engaged new voters. It is also clear that Corbyn engaged a particular set of younger, progressive voters like no recent party leader. We cannot assume it will apply next time to him or anyone else.
The power of the press is waning fast, replaced by social media
Unlike past elections, this wasn't noticeably affected by the tabloid press and their agenda. Despite helping to destroy the public images of Ed Miliband, Gordon Brown and other past Labour leaders, The Sun and Daily Mail's relentless two-year assault on Corbyn proved impotent.
Remarkably 30% of the former's readers voted for party led by a man they'd painted as terrorist-supporting, monarchy-bashing communist. And this among the declining, ageing segments that actually read newspapers. It is even possible that the mainstream media has a reverse effect nowadays - as each excess sparks viral outrage on social media, thus energising opponents.
Some said the 2015 result was all about the brilliant Tory digital campaign, particularly targeting voters on Facebook, but clearly Labour have more than caught up. Momentum's videos changed the conversation and overcame a monumental disparity in spending. This is now the key battleground.
Two-party politics is certainly not in irreversible decline
This was the ultimate re-alignment election, in the wake of Brexit, and the effect was the revival of two ancient, deeply polarised tribes. Having consistently fallen below 70%, the combined share between Labour and Conservatives topped 85% - it's highest level since 1970.
With tactical voting more prevalent than ever, people seem to be wising up to our electoral system. At the previous election, the Lib Dems were nearly wiped out in response to joining a Tory-led coalition and UKIP were rewarded with just one MP for their 4M votes. The old adage that backing smaller parties represents a wasted vote rang true.
Was this a one-off or will the trend persist? There are no shortage of divisions within each party and their supporters, particularly regarding Brexit. However the polling and parliamentary numbers are even bigger when discounting Scotland - where tactical voting against the SNP over independence has created an odd mixture of binary contests - and that inevitably means the voices of smaller parties will be further squeezed.