SCARCELY has a party been more disappointed with success.

For the past 17 years, Imran Khan’s Pakistan Tehreek-i-Insaf lingered in the political wilderness, only ever winning a single seat.

Now, the party has secured the second highest number of votes in the country. It will run a province, winning the highest number of seats there since the MMA’s vow to “revive fear of God”. It has seats in three provincial capitals and the federal capital. And it is the second most popular party in the two largest cities and the most populous province.

But instead of celebrating a modest triumph, Khan’s supporters have responded with paroxysms of sorrow and rage. Some can hardly believe they lost, and are quietly trying to swallow the indignity of defeat. Others don’t believe they lost, and are angrily denouncing the results as somehow less credible than Gen Zia’s 98.5pc referendum win.

In the tradition of Pakistani cricket captains, Khan promised them victory. And much like after a match that saw a few promising moments but ended in a crushing loss, many fans won’t accept an explanation that doesn’t maintain the outcome was fixed.

For over two years, Khan insisted he would be Pakistan’s next prime minister. The analogy of a flood, drowning opponents as it swept in, was quickly deemed inadequate. This was going to be a “tsunami”, he famously declared, with its violent waves destroying a despised political system.

The extravagant claims made some tactical sense. During an election campaign, no party says it will lose. To lure voters, Khan had to persuade them he was capable of winning. They wouldn’t have been tempted by expectations of third place, and five more years in purana Pakistan.

The mistake the PTI leadership made was that of a foolish army: it believed its own propaganda. On television, Khan advanced the complacent view that PTI would be swept to power by a wave of new young voters. No supportive data was furnished. Neither the media, nor Khan’s team, scrutinised the claim of a monolithic youth vote. In reality, young voters were divided.

In Lala Musa, for example, Qamar Zaman Kaira’s corner meetings featured a curious throwback to the 1970s, with teenagers chanting pro-PPP slogans in Punjabi. As polls now show, the bulk of Punjab’s youth voted for PML-N.

The PTI only gave polls convenient attention. When surveys of public opinion revealed them to be the most popular party, as they were for some months between end 2011 and early 2012, they breathlessly publicised the results. When the same polls showed them haemorrhaging support, it denounced them. The pollsters, they said, were in the pay of their N rivals and shouldn’t be taken seriously.

By contrast, N paid close attention to polling data. In Lahore, they surveyed key seats, and knew beforehand they would lose NA-126. In other constituencies, local MNAs commissioned their own polls, and then tried to overturn any negative perceptions.

There was also overdependence on the media. In PTI’s obscure years, Khan’s many television appearances yielded publicity disproportionate to his political clout. The exposure was crucial to the party’s recent growth. But when it came to a national campaign, airtime was a deceptive means of measuring popular support.

The media, keen for a competitive race, wasn’t going to spend six weeks talking about Nawaz Sharif cruising to power for a third time. With barely any campaigning in the three smaller provinces, television screens lent the illusion of a close race, with split screens showing Khan tirelessly gathering momentum with up to seven events a day in Punjab, while Sharif could only manage one or two there.

Jalsas and television ads, as the campaign showed, have a limited effect. Throughout South Asia, colourful rallies are key events. But they are only good at motivating an existing voter base. The sight of a leader rousing the party faithful might sway reluctant supporters. But rallies are a poor means of measuring support, or persuading new voters of a party’s worth.

A large rally in a city, where perhaps 50,000 people turn up, only represents a fraction of the total vote where each constituency has up to 400,000 registered voters. Khan’s aides, who tend to view their leader with unquestioning awe, would delight in assuring him that each successful jalsa represented a seat in the bag.

Television ads were good for news channels, some of which were able to pay off months of debts, but ultimately failed to shake the electorate.All of the negative ad campaigns failed, from the PPP’s swipes at Shahbaz Sharif to PML-N’s dig at Khan’s alliance with Sheikh Rasheed.Local efforts mattered more, where parties combined the clout of a viable candidate with a strong party ticket.

In Punjab, even strong PPP candidates collapsed under the oppressive weight of their ticket. The PTI ticket helped, as many respectable second place votes show, but victory proved elusive for obscure candidates. The PTI’s Punjab winners have all served in parliament before, or are related to former parliamentarians.

But PTI’s biggest mistake was targeting the wrong kind of voter. In KP, it tapped feelings of anti-incumbency and war-weariness. But in Punjab, it focused too narrowly on the thrusting but numerically small urban middle classes. It missed out on the poor majority. While PTI talked about visas and patwaris, PML-N offered those who can’t afford to travel abroad or sell land a seductively simple idea.

Khan conjured a fanciful dream of a new country the Swiss would envy. Sharif proved more effective in offering voters a more plausible return to an old country, where the lights work, fans whir, and shops do a reliable trade.

The writer covers Pakistan for TIME.