Room to move?

The mainstream media has championed the idea that the polls will narrow. They have said the polls would narrow after the budget, then it was after the election was called and now it is as the campaign progresses. That's great copy to maintain interest in proceedings for bewildered Liberal supporters, but the effect on the election depends on how many votes are available to cut Rudd's lead.

According to the pollsters, the accepted number of truly undecided voters going into a campaign is about 20-25%. Assume marginal seats, which by their nature have a higher percentage of undecided voters have between 25-30%.

Then look at Newspoll's marginal seat figure: July-Sep: ALP 58-42. Eagle-eyed readers would say, but hasn't Newspoll come back since then? The answer is no, not really. At worst, ALP would be 57-43 in the marginals, now trending better than the national average - that suggests possibly narrowing as 'true liberal' seats come back to the fold.

Assume then 56-44 is the national figure and 57-43 is the marginal figure:

Locked in preferences: National ALP 42 - 44.8, Lib/Nat 33 - 35.2
Marginal ALP 39.9 - 42.75 , Lib/Nat 30.1 - 32.25

If the national target for Lib/Nat to hang on is around 48 that means they need:

Either 15/25 or 12.8/20 - in other words, over 60% of undecided voters. This equates to Rudd putting in a consistently awful performance, slightly below Latham.

Even that might not save them. Remember the trajectory of Latham's campaign was one direction: down. All Rudd needs to do is bounce within two poles and he will deprive Howard of the momentum necessary to shift nearly two-thirds of undecided voters.

In the marginals, the Libs need somewhere near 50% to win,

19.9/30 or 17.75/25 - equating to two-thirds undecided voters. This would need something spectacular to go wrong, and it looks like the marginals are becoming proportionately better for the ALP, not worse.

If Rudd breaks even - we get ALP on 54.5 - 54.8 nationally and 54.9 - 55.25 in the marginals.

At 40% of marginal seat voters being undecided, Howard still needs 60% to reach 50% in those key seats. Given he traditionally doesn't make much a lot of ground in campaigns, losing votes in all of them except against Latham, that's a very tall order.

That would mean that the electorate would be so volatile you'd expect movements of 3-4% per poll. Most polls have barely moved that far in six months, indicating much less volatility.