In which I respond to Andrew's post responding to Claire column and add my thoughts on how history repeats... or not

I love viewing elections through the lens of history; the Americans do it so well and thoroughly, I wish there was more of it here. On that basis I really like what Claire Robertson's done in her Herald column – looking for patterns in the past. There's much truth to be found there. And I'm interested in the trend Claire has identified and her underlying point that most voters aren't as swingable as they might seem or say and many decide early.

It's a point strategists should wrestle with and learn from when they advise parties to keep their big policy announcements until the campaign. Arguably you have to lay the groundwork earlier.

But there is also the simple truth that all patterns are true... until they're not. It used to be that you had to win Eden to win an election... or Hamilton East (or was it West?). There are rules too about two term governments always getting third terms, but then you look at the specifics and see that situations that don't exist now – unprecedented prosperity, the Springbok Tour – were needed to get those governments over the line.

You see, nothing is ever quite the same and all these things pass. Every electoral rule gets broken at some point.

Andrew has thoroughly made some great points in response to Claire, but I can't help but want to add in my own thoughts, which agree with much of what he says, but some of Claire's points as well.

First, Andrew points out there are only six examples in Claire's data-set. I'd go further. There have been only two elections under MMP when we had parties seeking a third term, and you could make a case that they are only really comparable ones. What does each one tell us?

The first was 1996, which a) was the very first MMP election and b) was decided by Winston Peters's negotiations and deliberations rather than a clear mandate from voters, whether they decided how they'd vote a year out or a day out. Certainly a year out from the election few would have confidently predicted Peters would have chosen to support National, so it's hard to be clear just what voters knew or thought they were backing a year out.

The other was 2005. Look hard at that election and tell me campaigns don't matter. Student loans and the brethren moved the polls late; from memory a week out the swing was looking pretty good for Brash.

I tend to think Claire's right in the tension between Labour and the Greens. Even allowing for Andrew/Danyl's point that voters who stayed home last time could be vital, those voters will be mostly of the left and will be hotly contested by... Labour and the Greens. They could both grow their vote if they can get them out, but their public efforts to motivate them and win them to their side will likely create some hostility between the potential partners. Especially with National trying to paint the Greens as dragging Labour to extremes and Labour trying to win some from the centre as well.

In regards to the Maori Party, I'm back with Andrew. Flavell is less impressed with National than Sharples or Turia and could well be more interested in a coalition with Labour. Given the Tuhoe raids, unemployment and the like, the message he gets from his rohe could be very different from the one given to Turia or Sharples.

You could argue that the Maori Party maximised its power and attention by doing deals with National after the past two elections, but that it's now spent all the political capital it can on that trick and it needs to find another, such as switching sides to stay alive. Indeed, if you take Claire's approach and use history as your guide, you could say that the Maori Party has always followed power rather than personalities or eve  ideologies (as they always argue, they get most done for Maori when they're at the table) so if they could stay at the table and do so with Labour (where most of their party votes go), why assume they wouldn't?

And then there's New Zealand First. Again, Claire makes a good point about the cross-benches (or co-governance). But Peters is as likely to sit there propping up Labour as National, given his antipathy with Key and the policy gains he could win from Labour.

The other point that's new this term, and that Claire doesn't mention, is that the Greens are now a stable and distinct third party unlikely to drop back into single figures. The stability of its support was not a factor in the previous elections Claire is drawing from, but it will be next year. As Andrew and I have kept banging on, for the first time a two party bloc, even at a low ebb, has been only a few percentage points from overtaking the largest party. That's a new fact that could change the existing pattern.

So while nothing should ever be forgotten, it's also true that nothing is quite as it has been before. That also shouldn't be forgotten. History is a great guide, but it only allows us to see so far ahead, and even then the view can change as we walk.