Some Labour Party cheerleaders have convinced themselves they can capture the Treasury benches without winning an election. They're wrong.

They also seem to think Labour can afford to play nice with the Greens, play wordless footsies with Internet Mana, and avoid direct combat with National over the centre ground. Wrong, and wrong again.

This theory of September's election relies on the fantastical notion that a million-strong army of erstwhile non-voters (and, presumably, opinion poll non-responders) are set to storm the nation's polling booths once Labour has lurched exactly far enough left.

Such self-serving delusions are reminiscent of the American right's insistence prior to the 2012 elections that polls pointing to Obama's re-election were "skewed" against Republicans and Mitt Romney was heading for a landslide win.

Advertisement

Utterly convinced, as one pundit wrote, that voters "will send Obama into the history books as an undisputed calamity for America", even otherwise sensible conservatives latched on to the idea that older whites were under-sampled in pre-election surveys despite solid evidence that such voters are typically over-represented.

Confirmation bias - that most human tendency to believe whatever reinforces our existing beliefs - holds powerful sway, and especially among political partisans.

Comforting falsehoods keep supporters buoyed and dull the pain of imminent defeat, but they shouldn't take the place of a strategy.

The real world electoral equation in New Zealand remains unchanged from a century ago: hold your turf and plunder the middle.

In the First Past the Post era, the "middle" took the easily recognisable form of places like Gisborne, Horowhenua and Hamiltons East and West.

Labour had to compete in provincial and suburban marginal seats because forming a government wasn't possible in urban strongholds alone.

Some in the party took MMP as a cue to abandon marginal seat campaigning in favour of the party vote, diminishing Labour's geographic as well as demographic reach.

Inevitably, the party's standing took a hit in former battlegrounds: Labour, which topped the party vote in Hamilton West in 2005, won just 29 per cent last time.

Ironically for a party preoccupied with meeting diversity quotas, Labour's highly centralised approach has failed to produce a caucus notable for its dynamism or breadth of experience.

Would the recent proposal to ban trucks from fast lanes, for example, pass muster in a caucus with an assertive regional voice? It's highly doubtful.

Conversely, a policy agenda tilted towards regional voters risks smacking of opportunism this late in the electoral cycle, compounding the authenticity problem some perceive with Labour's leadership. Labour's posturing on immigration will backfire for that reason.

Cunliffe, Shearer, Parker or otherwise, no David can match Key's Goliath without competing across the spectrum: stemming the flow of millennial voters to the Greens; peeling swing voters off National in the regions and outer suburbs; and maximising so-called base voters in Labour's heartland.

The fear that such approaches are mutually exclusive - that appealing to the centre will invariably alienate the left and vice versa - reflects a deficiency of policy imagination and political confidence in roughly equal measure.

Labour is losing, and losing badly. And it's doing so for deep, structural reasons.

The party is too small, too monotonous, too narrowly focused.

A good start would be to revisit the fallacy that Labour gets to define the political terrain within which it competes. Hypothetical best-case scenarios built on complex MMP algorithms and phantom voter battalions are no match for a 20-point poll lead.

Phil Quin is a former adviser for the NZ and Australia labour parties and a strategic communications consultant.