Mike Hosking: Andrew Little's descent down a greasy poll

Potentially one of the big stars of this election season, which let's be honest as far as spectator sport goes has gotten off to a spectacular start, are the polls.

All three major polls this week had one consistency... and that was Labour's dismal standing at 24 per cent.

Now I could argue all day long how at 24 per cent Labour was still in it.
Labour was always going to need the Greens - and probably New Zealand First.
And that equation never really changed whether they were at 24 or 27.

Maybe in his mind Andrew Little had envisioned a much higher figure, a figure that really didn't require New Zealand First, a much simpler more reliable sort of arrangement, which would've meant he should've been well into the mid to high 30s by now - and when it became crushingly obvious that wasn't the case, he gave up.

I asked him last year when the traction was coming, and where he needed to be rating to be competitive ... he said they would be mid to high 30s by June/July.

Maybe all along that was his line in the sand. It doesn't matter now, he's gone.

To sum him up he seemed a decent bloke who tidied up the discipline in the party, but was never really more than the best of an ordinary lot who came out top in that weird leadership roadshow.

But apart from Labour's figures, what do we have from the polls? They are still volatile.

National are anywhere from 42 to 47.
That's government by a whisker, or not a hope in hell of a fourth term.

New Zealand First are anywhere from 11 to 16. For a smallish party, that is a massive difference - and it has implications not only as to whether they are even in the mix for government, but if they are, how many Cabinet seats they get, how much policy floats to the top, and how hard Winston asks to be prime minister.

The Greens' numbers are variable as well, so depending on which poll you cite, depends on whether you can argue Metiria Turei's mad foray into the world of illegality and fraud has paid dividends or not.

Polls of course enter this race with a bad name, but in many respects they shouldn't.
They have a bad name because of Brexit, the British election, and the presidential race in America.

The Brexit polls are the only ones that are instructive for us.
Why?

Because it was the only race where a national figure meant anything. The Brexit polls had a yay or nay number.

The same way under MMP 47 per cent equals 47 per cent of the seats.
In the British election and the US contest, there were no national races, and yet they rolled out national polls.

In Britain there were 650 individual races with a huge variety of possibilities and outcomes.

In the US election it was 50 state races, each with varied weightings as to what they contributed to the electoral college vote.

A national poll bore no resemblance to what was happening on the ground in any given state. And yet people having failed to grasp that, were then shocked when a generalised nationwide figure didn't dove-tail with what happened.

But what happened with Brexit (which was within the margin of error) - they all said it was a close race ... which it was.
As indeed is our one.
For National, 47 per cent is government. Or not. Both are possible.
At 46 you are almost certainly missing out.
At 48 you are in.

All figures are within the margin of error of any credible poll.
Now if you are going to blame a poll for a miss that small you don't understand polling.
It's a snapshot, it's a good guide without being a forensic map, or a sworn in blood promise.

Like the last three elections, this is a close run thing, and if you're reading the polls in a "keep the government " ... "change the government" sort of way - then that's exactly what they're telling us... and that won't change in the next 50-odd days.