Susan Hall “may win” fever swept Westminster like a whirlwind.
It sprung up within hours of the verification of votes starting at 9am on Friday in the mayoral contest.
Votes are not counted at this stage so the picture was still very sketchy.
But by then the parties had some intelligence on how the race had gone, based on tellers at polling stations, postal votes and early signs from the verification work, watched by party agents and some candidates, where votes are counted face up, to check they tally with the numbers from polling stations.
So it was plausible that the first reports, from Tory-leaning media, that Ms Hall was ahead in the Camden Barnet area could be true.
Labour had throughout the campaign insisted that the final result would be close, even though this was very much at odds with what the polls were saying.
But as the May 2 polling day approached, they closed and the Labour contender’s lead was halved, according to two of them.
With a margin of error of three per cent in surveys of 1,000 people, and a few more days for the polls to close further before polling day, the “close” election scenario was believeable.
Ms Hall might get a path to victory on low turn-out, if Mr Khan failed to squeeze the Green and Liberal Democrat vote, if the narrative that he is unpopular after two terms in City Hall proved true, if he lost support due to Labour’s position on Gaza, and if the switch to the first-past-the-post voting system ended up favouring the Tories.
Further reports from reliable sources suggested the Tory vote was holding up in parts of central London, but still with an expectation that Mr Khan would win, and that Ms Hall was doing well in parts of the Brent and Harrow constituency, her home turf.
By then, a political frenzy, fuelled by some Tory activists, was brewing about the possibility of Ms Hall pulling off a shock victory in one of the most catastrophic polling failures of all times.
Turnout figures published on Friday afternoon put rocket boosters under the febrile speculation.
They showed overall turnout slightly down, at 40.5 per cent, but far more interesting was the differential turnout.
In the City and East area, predominantly Labour, it was down to just over 31 per cent, but in the Tory stronghold of Bexley and Bromley over 48 per cent.
Could a huge army of “shy Tories”, furious at Mr Khan’s expansion of the Ultra Low Emission Zone to Outer London, be silently on the march to inflict a bombshell defeat on Labour?
Read More
Talk of London being hit by a “Sue-nami” spiralled, with speculation that such a victory for the Trump-supporting Tory candidate would transform the narrative of the elections for Rishi Sunak from a disaster to signs of recovery.
But it quickly became clear that the differential turnout large gap between the City and East/Bexley and Bromley was not being replicated across London.
These figures were also being over-interpreted as turnout in Inner London is often lower than in Outer London, and some of the central constituencies have bigger populations.
By now, previously nervous Labour sources were sounding more relaxed.
But astonishingly, the Susan Hall “may win” juggernaut was still picking up speed, with more and more claims that it would be close and that she was on course to beat Mr Khan.
Labour decided late afternoon on Friday to intervene to puncture the frenzy.
A source briefed that the fundamentals were good for Mr Khan, while still stressing that counting had still not started.
If Labour’s number crunchers were right, their man would secure a third term in office.
By Saturday morning, the Susan Hall “may win” political tornado had died down.
The actual result, Mr Khan gaining 43.8 per cent and Ms Hall 32.7 per cent, was almost exactly as predicted by the last Savanta poll of the campaign, for the Centre for London, published in the Evening Standard on May 1, putting the Labour candidate on 42 per cent and Tory challenger 32 per cent.
The political hysteria was over.
But it left in its wake searching questions over how it had managed to reach such a fever pitch and be so wrong.